remove claude tool use backwards compatability

feat(ollama): use official ollama library (#1859 )
Co-authored-by: Saoud Rizwan <7799382+saoudrizwan@users.noreply.github.com>
2026-09-14 10:41:33 +08:00 · 2025-03-04 21:49:48 -08:00 · 2025-03-04 14:46:26 -08:00 · 2025-03-03 22:11:44 -08:00 · 2025-03-03 20:06:42 -08:00 · 2025-03-03 20:01:53 -08:00
85 changed files with 3959 additions and 2531 deletions
@@ -0,0 +1,7 @@
+---
+"claude-dev": minor
+---
+Add Bedrock prompt caching support (optional).
+    
+This feature protected under checkbox because it is not yet rolled out to everyone, and if you will try to send cache headers, and its not enabled for you, you will get error.
+
@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Fixed problem with “win+shift+a” shortcut not working in Windows
@@ -1,5 +0,0 @@
---
-"claude-dev": minor
---
-
-Added X AI as a new provider with support for all current models including Grok-2 and Grok Vision. This integration enables users to connect to X AI's API using their API key and access models with context windows up to 131K tokens. The implementation includes proper handling for vision models and accurate pricing information.
@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Better Openrouter error typing and throwing more detailed messages
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-fix: make AWS Bedrock authentication predictable
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-Add translation to CODE_OF_CONDUCT, CONTRIBUTING and README to Arabic ar-sa.
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-A new MCP Marketplace display setting has been added to VSCode settings. (It is defaulted to "true")
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-Added a .clinerules file with details about our most common features
@@ -1,5 +0,0 @@
---
-"claude-dev": minor
---
-
-Added timeout configuration for individual MCP servers.
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-Make tags section in marketplace scrollable
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-opt: Enhance OpenAiHandler to diagnose request issues and handle empty streams
@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Add webview to lint job
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-.clineignore should be included in .gitignore
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-Migrate webview from CRA to Vite
@@ -1,5 +0,0 @@
---
-"claude-dev": minor
---
-
-Add rich MCP responses with images and link previews
@@ -1,5 +0,0 @@
---
-"claude-dev": minor
---
-
-Add preferred language option to settings
@@ -1,5 +0,0 @@
---
-"claude-dev": minor
---
-
-GPT 4.5 Preview added
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-Fix checking "Support Images" setting had no effect on first click
@@ -1,11 +0,0 @@
---
-"claude-dev": patch
---
-
-Add dynamic model fetching for the Requesty provider.
-
-Instead of manually typing the model name, the extension dynamically fetches
-all the supported model names from Requesty's /v1/models API.
-
-This allows users to use a fuzzy search logic when choosing the models and
-also guarantees the information for each model is up to date.
@@ -1,5 +0,0 @@
---
-"claude-dev": minor
---
-
-feat(vertex): Add prompt caching support for Claude on Vertex AI
@@ -1,11 +0,0 @@
---
-"claude-dev": minor
---
-
-   **Branch-Per-Task:** Each repo now has a single Shadow Git repo, with separate branches per task (instead of one Shadow Git repo per task).
-   **Legacy Support:** Existing Checkpoints remain functional, while all new Checkpoints use branch-per-task.
-
-   **Commits:** Legacy tasks commit to legacy Checkpoints; new tasks commit using branch-per-task.
-   **Diffing & Deletions:** Both legacy and branch-per-task Checkpoints support diffing and deletion.
-
-No migration needed—existing tasks stay as-is, and new tasks adopt **branch-per-task** automatically.
@@ -1,5 +0,0 @@
---
-"claude-dev": minor
---
-
-Update anthropic SDK to the latest version
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-fix: Anthropic's default model is 3.7
@@ -1,5 +0,0 @@
---
-"claude-dev": patch
---
-
-Adjustment to MCP server installation prompt
@@ -1,5 +0,0 @@
---
-"claude-dev": minor
---
-
-ADD IS_DEV and Hot Reloading to debug
@@ -1 +1 @@
-*       @saoudrizwan @ocasta181 @NightTrek @pashpashpash
+*       @saoudrizwan @ocasta181 @NightTrek @pashpashpash @dcbartlett
@@ -1,5 +1,24 @@
 # Changelog

+## [3.5.1]
+
+-   Add timeout option to MCP servers
+-   Add Gemini Flash models to Vertex provider (thanks @jpaodev!)
+-   Add prompt caching support for AWS Bedrock provider (thanks @buger!)
+-   Add AskSage provider (thanks @swhite24!)
+
+## [3.5.0]
+
+-   Add 'Enable extended thinking' option for Claude 3.7 Sonnet, with ability to set different budgets for Plan and Act modes
+-   Add support for rich MCP responses with automatic image previews, website thumbnails, and WolframAlpha visualizations
+-   Add language preference option in Advanced Settings
+-   Add xAI Provider Integration with support for all Grok models (thanks @andrewmonostate!)
+-   Fix issue with Linux XDG pointing to incorrect path for Document folder (thanks @jonatkinson!)
+
+## [3.4.10]
+
+-   Add support for GPT-4.5 preview model
+
 ## [3.4.9]

 -   Add toggle to let users opt-in to anonymous telemetry and error reporting
@@ -32,7 +32,7 @@ English | <a href="https://github.com/cline/cline/blob/main/locales/es/README.md

 Meet Cline, an AI assistant that can use your **CLI** a**N**d **E**ditor.

-Thanks to [Claude 3.5 Sonnet's agentic coding capabilities](https://www-cdn.anthropic.com/fed9cc193a14b84131812372d8d5857f8f304c52/Model_Card_Claude_3_Addendum.pdf), Cline can handle complex software development tasks step-by-step. With tools that let him create & edit files, explore large projects, use the browser, and execute terminal commands (after you grant permission), he can assist you in ways that go beyond code completion or tech support. Cline can even use the Model Context Protocol (MCP) to create new tools and extend his own capabilities. While autonomous AI scripts traditionally run in sandboxed environments, this extension provides a human-in-the-loop GUI to approve every file change and terminal command, providing a safe and accessible way to explore the potential of agentic AI.
+Thanks to [Claude 3.7 Sonnet's agentic coding capabilities](https://www.anthropic.com/claude/sonnet), Cline can handle complex software development tasks step-by-step. With tools that let him create & edit files, explore large projects, use the browser, and execute terminal commands (after you grant permission), he can assist you in ways that go beyond code completion or tech support. Cline can even use the Model Context Protocol (MCP) to create new tools and extend his own capabilities. While autonomous AI scripts traditionally run in sandboxed environments, this extension provides a human-in-the-loop GUI to approve every file change and terminal command, providing a safe and accessible way to explore the potential of agentic AI.

 1. Enter your task and add images to convert mockups into functional apps or fix bugs with screenshots.
 2. Cline starts by analyzing your file structure & source code ASTs, running regex searches, and reading relevant files to get up to speed in existing projects. By carefully managing what information is added to context, Cline can provide valuable assistance even for large, complex projects without overwhelming the context window.
@@ -1,3 +1,116 @@
-View our provacy policy on our website.
+# Cline Privacy Policy

-(Privacy Policy)[https://cline.bot/privacy]
+Cline Bot Inc. ("Cline," "we," "our," and/or "us") values the privacy of individuals who use our VS Code extension and related services (collectively, our "Services"). This privacy policy explains how we collect, use, and disclose information from users of our Services.
+
+## Key Points
+
+-   Cline operates entirely client-side as a VS Code extension
+-   No code or data is collected, stored, or transmitted to Cline's servers
+-   Your data is only sent to your chosen AI provider (e.g., Anthropic, OpenAI) when you explicitly request assistance
+-   All processing happens locally on your machine
+-   API keys are stored securely in VS Code's built-in settings storage
+
+## Information We Process
+
+### A. Information You Provide
+
+-   **API Keys**: When you choose to use certain AI model providers (OpenRouter, Anthropic, OpenAI, etc.), you provide API keys. These are stored securely and locally in your VS Code settings.
+-   **Communications**: If you contact us directly (e.g., via Discord or email), we may receive information like your name, email address, and message contents.
+
+### B. Information Processing
+
+Cline functions solely as a client-side VS Code extension that facilitates communication between your editor and your chosen AI model provider:
+
+1. **File Contents**:
+
+    - Only sent to your chosen AI provider when you explicitly request assistance
+    - Never stored or transmitted to Cline's servers
+    - Only the specific files/content you select are included
+
+2. **Terminal Commands**:
+
+    - Processed entirely locally on your machine
+    - Require explicit user confirmation before execution
+    - No command history is transmitted to Cline
+
+3. **Browser Integration**:
+    - Screenshots and console logs are processed locally
+    - Temporary data is cleared after task completion
+
+## Data Security
+
+1. **Local-Only Processing**:
+
+    - All operations happen on your local machine
+    - No central servers or data collection by default
+    - Anonymous telemetry and usage statistics are only collected if you explicitly opt in
+    - No account creation required
+
+2. **API Key Security**:
+
+    - Stored using VS Code's secure settings storage system
+    - Never transmitted to Cline's servers
+    - You can remove/modify keys at any time
+
+3. **User Control**:
+    - Explicit approval required for file changes
+    - Terminal commands require confirmation
+    - Browser actions need explicit permission
+    - You control which AI provider to use
+
+## Communication with AI Providers
+
+When you request assistance:
+
+1. Selected content is sent directly to your chosen AI provider
+2. No data passes through Cline's servers
+3. Provider's own privacy policy applies to this communication:
+    - [Anthropic Privacy Policy](https://www.anthropic.com/privacy)
+    - [OpenAI Privacy Policy](https://openai.com/privacy)
+    - [OpenRouter Privacy Policy](https://openrouter.ai/privacy)
+
+## Error Handling & Debugging
+
+-   Error logs are processed locally
+-   No automatic error reporting to Cline
+    - Optional anonymous telemetry and error reporting via PostHog if you opt in
+-   You control what information to include when manually reporting issues
+
+## Children's Privacy
+
+We do not knowingly collect, maintain, or use personal information from children under 18 years of age, and no part of our Service(s) is directed to children. If you learn that a child has provided us with personal information in violation of this Privacy Policy, then you may alert us at support@cline.bot.
+
+## Changes to Privacy Policy
+
+We will post any changes to this policy on our GitHub repository. Significant changes will be announced in our Discord community.
+
+## Security Concerns & Auditing
+
+-   Cline is open source and available for security audit
+-   Our client-side architecture ensures no central point of data collection
+-   You can inspect exactly what data is being sent to AI providers
+-   Enterprise users can implement additional access controls through VS Code
+
+## Telemetry & Usage Statistics
+
+If you choose to opt in to anonymous telemetry:
+
+-   Basic usage statistics and error reports are collected via PostHog
+-   A stable, anonymous identifier (VS Code's `machineId`) is used to understand unique usage patterns
+    - This identifier is not linked to any personal information
+    - It helps us understand how features are used across sessions
+    - It cannot be used to identify you personally
+-   All data is anonymized and cannot be linked to individual users
+-   No code content or sensitive information is ever included
+-   You can opt out at any time through:
+    - VS Code Settings > Cline > Enable Telemetry
+    - VS Code Settings > Telemetry > Telemetry Level (setting this to anything other than "all" will disable Cline's telemetry)
+-   Collected data helps us improve the extension's functionality and stability
+
+## Contact Us
+
+For privacy-related questions or concerns:
+
+-   Open an issue on our [GitHub repository](https://github.com/cline/cline)
+-   Join our [Discord community](https://discord.gg/cline)
+-   Email: support@cline.bot
@@ -1,3 +0,0 @@
-View our Terms of Service on our website.
-
-(Terms of Service)[https://cline.bot/tos]
@@ -32,7 +32,7 @@

 التقى Cline، مساعد الذكاء الاصطناعي الذي يمكنه استخدام **سطر الأوامر** و **محرر النصوص** الخاص بك.

-بفضل [قدرات Claude 3.5 Sonnet على التعليمات البرمجية الوكيلة](https://www-cdn.anthropic.com/fed9cc193a14b84131812372d8d5857f8f304c52/Model_Card_Claude_3_Addendum.pdf)، يمكن لـ Cline التعامل مع مهام تطوير البرامج المعقدة خطوة بخطوة. مع الأدوات التي تسمح له بإنشاء وتعديل الملفات، واستكشاف المشاريع الكبيرة، واستخدام المتصفح، وتنفيذ أوامر الطرفية (بعد منحك الإذن)، يمكنه مساعدتك بطرق تتجاوز إكمال الكود أو الدعم الفني. يمكن لـ Cline أيضًا استخدام بروتوكول سياق النموذج (MCP) لإنشاء أدوات جديدة وتوسيع قدراته الخاصة. في حين تعمل النصوص البرمجية الآلية المستقلة تقليديًا في بيئات محاصرة، توفر هذه الإضافة واجهة رسومية لموافقة المستخدم على كل تغيير في الملف وأمر طرفية، مما يوفر طريقة آمنة وسهلة الاستخدام لاستكشاف إمكانات الذكاء الاصطناعي الوكيل.
+بفضل [قدرات Claude 3.7 Sonnet على التعليمات البرمجية الوكيلة](https://www.anthropic.com/claude/sonnet)، يمكن لـ Cline التعامل مع مهام تطوير البرامج المعقدة خطوة بخطوة. مع الأدوات التي تسمح له بإنشاء وتعديل الملفات، واستكشاف المشاريع الكبيرة، واستخدام المتصفح، وتنفيذ أوامر الطرفية (بعد منحك الإذن)، يمكنه مساعدتك بطرق تتجاوز إكمال الكود أو الدعم الفني. يمكن لـ Cline أيضًا استخدام بروتوكول سياق النموذج (MCP) لإنشاء أدوات جديدة وتوسيع قدراته الخاصة. في حين تعمل النصوص البرمجية الآلية المستقلة تقليديًا في بيئات محاصرة، توفر هذه الإضافة واجهة رسومية لموافقة المستخدم على كل تغيير في الملف وأمر طرفية، مما يوفر طريقة آمنة وسهلة الاستخدام لاستكشاف إمكانات الذكاء الاصطناعي الوكيل.

 1. أدخل مهمتك وأضف الصور لتحويل المحاكاة إلى تطبيقات وظيفية أو إصلاح الأخطاء مع لقطات الشاشة.
 2. يبدأ Cline بتحليل هيكل الملفات الخاصة بك وشجرة التعريف المصدرية، وإجراء عمليات بحث regex، وقراءة الملفات ذات الصلة للاطلاع على المشاريع الحالية. من خلال إدارة المعلومات التي يتم إضافتها إلى السياق بعناية، يمكن لـ Cline تقديم مساعدة قيمة حتى للمشاريع الكبيرة والمعقدة دون إرهاق نافذة السياق.
@@ -28,7 +28,7 @@

 Lernen Sie Cline kennen, einen KI-Assistenten, der Ihre **CLI** u**N**d **E**ditor nutzen kann.

-Dank der [agentischen Codierungsfähigkeiten von Claude 3.5 Sonnet](https://www-cdn.anthropic.com/fed9cc193a14b84131812372d8d5857f8f304c52/Model_Card_Claude_3_Addendum.pdf) kann Cline komplexe Softwareentwicklungsaufgaben Schritt für Schritt bewältigen. Mit Werkzeugen, die ihm das Erstellen und Bearbeiten von Dateien, das Erkunden großer Projekte, die Nutzung des Browsers und das Ausführen von Terminalbefehlen (nach Ihrer Genehmigung) ermöglichen, kann er Ihnen auf eine Weise helfen, die über die Codevervollständigung oder technischen Support hinausgeht. Cline kann sogar das Model Context Protocol (MCP) verwenden, um neue Werkzeuge zu erstellen und seine eigenen Fähigkeiten zu erweitern. Während autonome KI-Skripte traditionell in sandboxed Umgebungen laufen, bietet diese Erweiterung eine Mensch-in-der-Schleife-GUI, um jede Dateiänderung und jeden Terminalbefehl zu genehmigen, was eine sichere und zugängliche Möglichkeit bietet, das Potenzial agentischer KI zu erkunden.
+Dank der [agentischen Codierungsfähigkeiten von Claude 3.7 Sonnet](https://www.anthropic.com/claude/sonnet) kann Cline komplexe Softwareentwicklungsaufgaben Schritt für Schritt bewältigen. Mit Werkzeugen, die ihm das Erstellen und Bearbeiten von Dateien, das Erkunden großer Projekte, die Nutzung des Browsers und das Ausführen von Terminalbefehlen (nach Ihrer Genehmigung) ermöglichen, kann er Ihnen auf eine Weise helfen, die über die Codevervollständigung oder technischen Support hinausgeht. Cline kann sogar das Model Context Protocol (MCP) verwenden, um neue Werkzeuge zu erstellen und seine eigenen Fähigkeiten zu erweitern. Während autonome KI-Skripte traditionell in sandboxed Umgebungen laufen, bietet diese Erweiterung eine Mensch-in-der-Schleife-GUI, um jede Dateiänderung und jeden Terminalbefehl zu genehmigen, was eine sichere und zugängliche Möglichkeit bietet, das Potenzial agentischer KI zu erkunden.

 1. Geben Sie Ihre Aufgabe ein und fügen Sie Bilder hinzu, um Mockups in funktionale Apps zu konvertieren oder Fehler mit Screenshots zu beheben.
 2. Cline beginnt mit der Analyse Ihrer Dateistruktur und Quellcode-ASTs, führt Regex-Suchen durch und liest relevante Dateien, um sich in bestehenden Projekten zurechtzufinden. Durch sorgfältiges Management der hinzugefügten Informationen kann Cline wertvolle Unterstützung auch bei großen, komplexen Projekten bieten, ohne das Kontextfenster zu überladen.
@@ -28,7 +28,7 @@

 Conozca a Cline, un asistente de IA que puede usar su **CLI** y **E**ditor.

-Gracias a las [habilidades de codificación agencial de Claude 3.5 Sonnet](https://www-cdn.anthropic.com/fed9cc193a14b84131812372d8d5857f8f304c52/Model_Card_Claude_3_Addendum.pdf), Cline puede abordar tareas complejas de desarrollo de software paso a paso. Con herramientas que le permiten crear y editar archivos, explorar grandes proyectos, usar el navegador y ejecutar comandos de terminal (con su aprobación), puede ayudarle de una manera que va más allá de la autocompletación de código o el soporte técnico. Cline incluso puede usar el Model Context Protocol (MCP) para crear nuevas herramientas y expandir sus propias capacidades. Mientras que los scripts de IA autónomos tradicionalmente se ejecutan en entornos aislados, esta extensión ofrece una GUI con un humano en el bucle para aprobar cada cambio de archivo y comando de terminal, proporcionando una forma segura y accesible de explorar el potencial de la IA agencial.
+Gracias a las [habilidades de codificación agencial de Claude 3.7 Sonnet](https://www.anthropic.com/claude/sonnet), Cline puede abordar tareas complejas de desarrollo de software paso a paso. Con herramientas que le permiten crear y editar archivos, explorar grandes proyectos, usar el navegador y ejecutar comandos de terminal (con su aprobación), puede ayudarle de una manera que va más allá de la autocompletación de código o el soporte técnico. Cline incluso puede usar el Model Context Protocol (MCP) para crear nuevas herramientas y expandir sus propias capacidades. Mientras que los scripts de IA autónomos tradicionalmente se ejecutan en entornos aislados, esta extensión ofrece una GUI con un humano en el bucle para aprobar cada cambio de archivo y comando de terminal, proporcionando una forma segura y accesible de explorar el potencial de la IA agencial.

 1. Ingrese su tarea y agregue imágenes para convertir maquetas en aplicaciones funcionales o solucionar errores con capturas de pantalla.
 2. Cline comenzará analizando su estructura de archivos y ASTs de código fuente, realizando búsquedas Regex y leyendo archivos relevantes para orientarse en proyectos existentes. Al gestionar cuidadosamente la información agregada, Cline puede proporcionar asistencia valiosa incluso en proyectos grandes y complejos sin sobrecargar la ventana de contexto.
@@ -28,7 +28,7 @@

 Clineは、**CLI**と**エディター**を使用できるAIアシスタントです。

-[Claude 3.5 Sonnetのエージェント的コーディング機能](https://www-cdn.anthropic.com/fed9cc193a14b84131812372d8d5857f8f304c52/Model_Card_Claude_3_Addendum.pdf)のおかげで、Clineは複雑なソフトウェア開発タスクをステップバイステップで処理できます。ファイルの作成と編集、大規模プロジェクトの探索、ブラウザの使用、ターミナルコマンドの実行（許可後）などのツールを使用して、コード補完や技術サポートを超えた支援を提供します。Clineは、Model Context Protocol (MCP)を使用して新しいツールを作成し、自身の機能を拡張することもできます。自律的なAIスクリプトは通常サンドボックス環境で実行されますが、この拡張機能はファイル変更やターミナルコマンドを承認するための人間インターフェースを提供し、エージェント的AIの可能性を安全かつアクセスしやすい方法で探求できます。
+[Claude 3.7 Sonnetのエージェント的コーディング機能](https://www.anthropic.com/claude/sonnet)のおかげで、Clineは複雑なソフトウェア開発タスクをステップバイステップで処理できます。ファイルの作成と編集、大規模プロジェクトの探索、ブラウザの使用、ターミナルコマンドの実行（許可後）などのツールを使用して、コード補完や技術サポートを超えた支援を提供します。Clineは、Model Context Protocol (MCP)を使用して新しいツールを作成し、自身の機能を拡張することもできます。自律的なAIスクリプトは通常サンドボックス環境で実行されますが、この拡張機能はファイル変更やターミナルコマンドを承認するための人間インターフェースを提供し、エージェント的AIの可能性を安全かつアクセスしやすい方法で探求できます。

 1. タスクを入力し、モックアップを機能するアプリに変換したり、スクリーンショットでバグを修正したりします。
 2. Clineは、ファイル構造とソースコードASTの分析、正規表現検索の実行、関連ファイルの読み取りから始め、既存プロジェクトに精通します。コンテキストに追加される情報を慎重に管理することで、大規模で複雑なプロジェクトでもコンテキストウィンドウを圧倒することなく貴重な支援を提供できます。
@@ -28,7 +28,7 @@

 Conheça o Cline: um assistente de IA que pode usar seu **CLI** e **Editor**.

-Graças às [habilidades avançadas do Claude 3.5 Sonnet](https://www-cdn.anthropic.com/fed9cc193a14b84131812372d8d5857f8f304c52/Model_Card_Claude_3_Addendum.pdf), o Cline pode lidar com tarefas complexas de desenvolvimento de software passo a passo. Com ferramentas que permitem criar e editar arquivos, explorar grandes projetos, usar o navegador e executar comandos no terminal (com sua aprovação), ele pode ajudar você de maneiras que vão além da inclusão de código ou suporte técnico. O Cline pode é capaz inclusive de usar o Model Context Protocol (MCP) para criar novas ferramentas e expandir seus próprios recursos. Embora os scripts de IA autônomas tradicionalmente sejam executados em ambientes isolados, esta extensão oferece uma GUI com um humano no circuito para aprovar cada alteração de arquivo e comando de terminal, fornecendo uma maneira segura e acessível de explorar todo o potencial da IA.
+Graças às [habilidades avançadas do Claude 3.7 Sonnet](https://www.anthropic.com/claude/sonnet), o Cline pode lidar com tarefas complexas de desenvolvimento de software passo a passo. Com ferramentas que permitem criar e editar arquivos, explorar grandes projetos, usar o navegador e executar comandos no terminal (com sua aprovação), ele pode ajudar você de maneiras que vão além da inclusão de código ou suporte técnico. O Cline pode é capaz inclusive de usar o Model Context Protocol (MCP) para criar novas ferramentas e expandir seus próprios recursos. Embora os scripts de IA autônomas tradicionalmente sejam executados em ambientes isolados, esta extensão oferece uma GUI com um humano no circuito para aprovar cada alteração de arquivo e comando de terminal, fornecendo uma maneira segura e acessível de explorar todo o potencial da IA.

 1. Insira sua tarefa e adicione imagens para transformar mockups em aplicativos funcionais ou corrigir erros através de capturas de tela.

@@ -28,7 +28,7 @@

 认识 Cline，一个可以使用你的 **CLI** 和 **编辑器** 的 AI 助手。

-感谢 [Claude 3.5 Sonnet 的代理编码能力](https://www-cdn.anthropic.com/fed9cc193a14b84131812372d8d5857f8f304c52/Model_Card_Claude_3_Addendum.pdf)，Cline 可以一步步处理复杂的软件开发任务。通过允许他创建和编辑文件、探索大型项目、使用浏览器和执行终端命令（在你授予权限后），他可以提供超越代码完成或技术支持的帮助。Cline 甚至可以使用 Model Context Protocol (MCP) 创建新工具并扩展自己的能力。虽然自主 AI 脚本传统上在沙盒环境中运行，但此扩展提供了一个人机交互的 GUI 来批准每个文件更改和终端命令，提供了一种安全且可访问的方式来探索代理 AI 的潜力。
+感谢 [Claude 3.7 Sonnet 的代理编码能力](https://www.anthropic.com/claude/sonnet)，Cline 可以一步步处理复杂的软件开发任务。通过允许他创建和编辑文件、探索大型项目、使用浏览器和执行终端命令（在你授予权限后），他可以提供超越代码完成或技术支持的帮助。Cline 甚至可以使用 Model Context Protocol (MCP) 创建新工具并扩展自己的能力。虽然自主 AI 脚本传统上在沙盒环境中运行，但此扩展提供了一个人机交互的 GUI 来批准每个文件更改和终端命令，提供了一种安全且可访问的方式来探索代理 AI 的潜力。

 1. 输入你的任务并添加图像，将模型转换为功能应用程序或通过截图修复错误。
 2. Cline 首先分析你的文件结构和源代码 AST，运行正则表达式搜索，并阅读相关文件以了解现有项目。通过仔细管理添加到上下文中的信息，Cline 即使在大型复杂项目中也能提供有价值的帮助，而不会使上下文窗口过载。
@@ -28,7 +28,7 @@

 認識 Cline，一個可以使用你的 **CLI** 和 **編輯器** 的 AI 助手。

-感謝 [Claude 3.5 Sonnet 的代理編碼能力](https://www-cdn.anthropic.com/fed9cc193a14b84131812372d8d5857f8f304c52/Model_Card_Claude_3_Addendum.pdf)，Cline 可以一步步處理複雜的軟件開發任務。通過允許他創建和編輯文件、探索大型項目、使用瀏覽器和執行終端命令（在你授予權限後），他可以提供超越代碼完成或技術支持的幫助。Cline 甚至可以使用 Model Context Protocol (MCP) 創建新工具並擴展自己的能力。雖然自主 AI 腳本傳統上在沙盒環境中運行，但此擴展提供了一個人機交互的 GUI 來批准每個文件更改和終端命令，提供了一種安全且可訪問的方式來探索代理 AI 的潛力。
+感謝 [Claude 3.7 Sonnet 的代理編碼能力](https://www.anthropic.com/claude/sonnet)，Cline 可以一步步處理複雜的軟件開發任務。通過允許他創建和編輯文件、探索大型項目、使用瀏覽器和執行終端命令（在你授予權限後），他可以提供超越代碼完成或技術支持的幫助。Cline 甚至可以使用 Model Context Protocol (MCP) 創建新工具並擴展自己的能力。雖然自主 AI 腳本傳統上在沙盒環境中運行，但此擴展提供了一個人機交互的 GUI 來批准每個文件更改和終端命令，提供了一種安全且可訪問的方式來探索代理 AI 的潛力。

 1. 輸入你的任務並添加圖像，將模型轉換為功能應用程序或通過截圖修復錯誤。
 2. Cline 首先分析你的文件結構和源代碼 AST，運行正則表達式搜索，並閱讀相關文件以了解現有項目。通過仔細管理添加到上下文中的信息，Cline 即使在大型複雜項目中也能提供有價值的幫助，而不會使上下文窗口過載。
@@ -1,17 +1,18 @@
 {
 	"name": "claude-dev",
-	"version": "3.4.9",
+	"version": "3.5.1",
 	"lockfileVersion": 3,
 	"requires": true,
 	"packages": {
 		"": {
 			"name": "claude-dev",
-			"version": "3.4.9",
+			"version": "3.5.1",
 			"license": "Apache-2.0",
 			"dependencies": {
 				"@anthropic-ai/bedrock-sdk": "^0.12.4",
 				"@anthropic-ai/sdk": "^0.37.0",
 				"@anthropic-ai/vertex-sdk": "^0.6.4",
+				"@google-cloud/vertexai": "^1.9.3",
 				"@google/generative-ai": "^0.18.0",
 				"@mistralai/mistralai": "^1.5.0",
 				"@modelcontextprotocol/sdk": "^1.0.1",
@@ -36,6 +37,7 @@
 				"isbinaryfile": "^5.0.2",
 				"mammoth": "^1.8.0",
 				"monaco-vscode-textmate-theme-converter": "^0.1.7",
+				"ollama": "^0.5.13",
 				"open-graph-scraper": "^6.9.0",
 				"openai": "^4.83.0",
 				"os-name": "^6.0.0",
@@ -112,9 +114,9 @@
 			}
 		},
 		"node_modules/@anthropic-ai/sdk/node_modules/@types/node": {
-			"version": "18.19.76",
-			"resolved": "https://registry.npmjs.org/@types/node/-/node-18.19.76.tgz",
-			"integrity": "sha512-yvR7Q9LdPz2vGpmpJX5LolrgRdWvB67MJKDPSgIIzpFbaf9a1j/f5DnLp5VDyHGMR0QZHlTr1afsD87QCXFHKw==",
+			"version": "18.19.78",
+			"resolved": "https://registry.npmjs.org/@types/node/-/node-18.19.78.tgz",
+			"integrity": "sha512-m1ilZCTwKLkk9rruBJXFeYN0Bc5SbjirwYX/Td3MqPfioYbgun3IvK/m8dQxMCnrPGZPg1kvXjp3SIekCN/ynw==",
 			"license": "MIT",
 			"dependencies": {
 				"undici-types": "~5.26.4"
@@ -130,6 +132,30 @@
 				"google-auth-library": "^9.4.2"
 			}
 		},
+		"node_modules/@anthropic-ai/vertex-sdk/node_modules/@anthropic-ai/sdk": {
+			"version": "0.38.0",
+			"resolved": "https://registry.npmjs.org/@anthropic-ai/sdk/-/sdk-0.38.0.tgz",
+			"integrity": "sha512-ZUYjadEEeb1wwKd9MEM+plSfl7zQjYixhjHRtyPsjO7MtzRmbZvBb1n1ofo2kHwb0aUiIdb3aAEbAwS9Bcbm/A==",
+			"license": "MIT",
+			"dependencies": {
+				"@types/node": "^18.11.18",
+				"@types/node-fetch": "^2.6.4",
+				"abort-controller": "^3.0.0",
+				"agentkeepalive": "^4.2.1",
+				"form-data-encoder": "1.7.2",
+				"formdata-node": "^4.3.2",
+				"node-fetch": "^2.6.7"
+			}
+		},
+		"node_modules/@anthropic-ai/vertex-sdk/node_modules/@types/node": {
+			"version": "18.19.76",
+			"resolved": "https://registry.npmjs.org/@types/node/-/node-18.19.76.tgz",
+			"integrity": "sha512-yvR7Q9LdPz2vGpmpJX5LolrgRdWvB67MJKDPSgIIzpFbaf9a1j/f5DnLp5VDyHGMR0QZHlTr1afsD87QCXFHKw==",
+			"license": "MIT",
+			"dependencies": {
+				"undici-types": "~5.26.4"
+			}
+		},
 		"node_modules/@aws-crypto/crc32": {
 			"version": "5.2.0",
 			"resolved": "https://registry.npmjs.org/@aws-crypto/crc32/-/crc32-5.2.0.tgz",
@@ -2538,6 +2564,74 @@
 				"url": "https://github.com/prettier/prettier?sponsor=1"
 			}
 		},
+		"node_modules/@esbuild/aix-ppc64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/aix-ppc64/-/aix-ppc64-0.25.0.tgz",
+			"integrity": "sha512-O7vun9Sf8DFjH2UtqK8Ku3LkquL9SZL8OLY1T5NZkA34+wG3OQF7cl4Ql8vdNzM6fzBbYfLaiRLIOZ+2FOCgBQ==",
+			"cpu": [
+				"ppc64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"aix"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/android-arm": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/android-arm/-/android-arm-0.25.0.tgz",
+			"integrity": "sha512-PTyWCYYiU0+1eJKmw21lWtC+d08JDZPQ5g+kFyxP0V+es6VPPSUhM6zk8iImp2jbV6GwjX4pap0JFbUQN65X1g==",
+			"cpu": [
+				"arm"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"android"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/android-arm64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/android-arm64/-/android-arm64-0.25.0.tgz",
+			"integrity": "sha512-grvv8WncGjDSyUBjN9yHXNt+cq0snxXbDxy5pJtzMKGmmpPxeAmAhWxXI+01lU5rwZomDgD3kJwulEnhTRUd6g==",
+			"cpu": [
+				"arm64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"android"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/android-x64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/android-x64/-/android-x64-0.25.0.tgz",
+			"integrity": "sha512-m/ix7SfKG5buCnxasr52+LI78SQ+wgdENi9CqyCXwjVR2X4Jkz+BpC3le3AoBPYTC9NHklwngVXvbJ9/Akhrfg==",
+			"cpu": [
+				"x64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"android"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
 		"node_modules/@esbuild/darwin-arm64": {
 			"version": "0.25.0",
 			"resolved": "https://registry.npmjs.org/@esbuild/darwin-arm64/-/darwin-arm64-0.25.0.tgz",
@@ -2555,6 +2649,346 @@
 				"node": ">=18"
 			}
 		},
+		"node_modules/@esbuild/darwin-x64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/darwin-x64/-/darwin-x64-0.25.0.tgz",
+			"integrity": "sha512-DgDaYsPWFTS4S3nWpFcMn/33ZZwAAeAFKNHNa1QN0rI4pUjgqf0f7ONmXf6d22tqTY+H9FNdgeaAa+YIFUn2Rg==",
+			"cpu": [
+				"x64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"darwin"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/freebsd-arm64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/freebsd-arm64/-/freebsd-arm64-0.25.0.tgz",
+			"integrity": "sha512-VN4ocxy6dxefN1MepBx/iD1dH5K8qNtNe227I0mnTRjry8tj5MRk4zprLEdG8WPyAPb93/e4pSgi1SoHdgOa4w==",
+			"cpu": [
+				"arm64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"freebsd"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/freebsd-x64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/freebsd-x64/-/freebsd-x64-0.25.0.tgz",
+			"integrity": "sha512-mrSgt7lCh07FY+hDD1TxiTyIHyttn6vnjesnPoVDNmDfOmggTLXRv8Id5fNZey1gl/V2dyVK1VXXqVsQIiAk+A==",
+			"cpu": [
+				"x64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"freebsd"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/linux-arm": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/linux-arm/-/linux-arm-0.25.0.tgz",
+			"integrity": "sha512-vkB3IYj2IDo3g9xX7HqhPYxVkNQe8qTK55fraQyTzTX/fxaDtXiEnavv9geOsonh2Fd2RMB+i5cbhu2zMNWJwg==",
+			"cpu": [
+				"arm"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"linux"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/linux-arm64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/linux-arm64/-/linux-arm64-0.25.0.tgz",
+			"integrity": "sha512-9QAQjTWNDM/Vk2bgBl17yWuZxZNQIF0OUUuPZRKoDtqF2k4EtYbpyiG5/Dk7nqeK6kIJWPYldkOcBqjXjrUlmg==",
+			"cpu": [
+				"arm64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"linux"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/linux-ia32": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/linux-ia32/-/linux-ia32-0.25.0.tgz",
+			"integrity": "sha512-43ET5bHbphBegyeqLb7I1eYn2P/JYGNmzzdidq/w0T8E2SsYL1U6un2NFROFRg1JZLTzdCoRomg8Rvf9M6W6Gg==",
+			"cpu": [
+				"ia32"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"linux"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/linux-loong64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/linux-loong64/-/linux-loong64-0.25.0.tgz",
+			"integrity": "sha512-fC95c/xyNFueMhClxJmeRIj2yrSMdDfmqJnyOY4ZqsALkDrrKJfIg5NTMSzVBr5YW1jf+l7/cndBfP3MSDpoHw==",
+			"cpu": [
+				"loong64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"linux"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/linux-mips64el": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/linux-mips64el/-/linux-mips64el-0.25.0.tgz",
+			"integrity": "sha512-nkAMFju7KDW73T1DdH7glcyIptm95a7Le8irTQNO/qtkoyypZAnjchQgooFUDQhNAy4iu08N79W4T4pMBwhPwQ==",
+			"cpu": [
+				"mips64el"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"linux"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/linux-ppc64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/linux-ppc64/-/linux-ppc64-0.25.0.tgz",
+			"integrity": "sha512-NhyOejdhRGS8Iwv+KKR2zTq2PpysF9XqY+Zk77vQHqNbo/PwZCzB5/h7VGuREZm1fixhs4Q/qWRSi5zmAiO4Fw==",
+			"cpu": [
+				"ppc64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"linux"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/linux-riscv64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/linux-riscv64/-/linux-riscv64-0.25.0.tgz",
+			"integrity": "sha512-5S/rbP5OY+GHLC5qXp1y/Mx//e92L1YDqkiBbO9TQOvuFXM+iDqUNG5XopAnXoRH3FjIUDkeGcY1cgNvnXp/kA==",
+			"cpu": [
+				"riscv64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"linux"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/linux-s390x": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/linux-s390x/-/linux-s390x-0.25.0.tgz",
+			"integrity": "sha512-XM2BFsEBz0Fw37V0zU4CXfcfuACMrppsMFKdYY2WuTS3yi8O1nFOhil/xhKTmE1nPmVyvQJjJivgDT+xh8pXJA==",
+			"cpu": [
+				"s390x"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"linux"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/linux-x64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/linux-x64/-/linux-x64-0.25.0.tgz",
+			"integrity": "sha512-9yl91rHw/cpwMCNytUDxwj2XjFpxML0y9HAOH9pNVQDpQrBxHy01Dx+vaMu0N1CKa/RzBD2hB4u//nfc+Sd3Cw==",
+			"cpu": [
+				"x64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"linux"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/netbsd-arm64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/netbsd-arm64/-/netbsd-arm64-0.25.0.tgz",
+			"integrity": "sha512-RuG4PSMPFfrkH6UwCAqBzauBWTygTvb1nxWasEJooGSJ/NwRw7b2HOwyRTQIU97Hq37l3npXoZGYMy3b3xYvPw==",
+			"cpu": [
+				"arm64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"netbsd"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/netbsd-x64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/netbsd-x64/-/netbsd-x64-0.25.0.tgz",
+			"integrity": "sha512-jl+qisSB5jk01N5f7sPCsBENCOlPiS/xptD5yxOx2oqQfyourJwIKLRA2yqWdifj3owQZCL2sn6o08dBzZGQzA==",
+			"cpu": [
+				"x64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"netbsd"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/openbsd-arm64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/openbsd-arm64/-/openbsd-arm64-0.25.0.tgz",
+			"integrity": "sha512-21sUNbq2r84YE+SJDfaQRvdgznTD8Xc0oc3p3iW/a1EVWeNj/SdUCbm5U0itZPQYRuRTW20fPMWMpcrciH2EJw==",
+			"cpu": [
+				"arm64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"openbsd"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/openbsd-x64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/openbsd-x64/-/openbsd-x64-0.25.0.tgz",
+			"integrity": "sha512-2gwwriSMPcCFRlPlKx3zLQhfN/2WjJ2NSlg5TKLQOJdV0mSxIcYNTMhk3H3ulL/cak+Xj0lY1Ym9ysDV1igceg==",
+			"cpu": [
+				"x64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"openbsd"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/sunos-x64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/sunos-x64/-/sunos-x64-0.25.0.tgz",
+			"integrity": "sha512-bxI7ThgLzPrPz484/S9jLlvUAHYMzy6I0XiU1ZMeAEOBcS0VePBFxh1JjTQt3Xiat5b6Oh4x7UC7IwKQKIJRIg==",
+			"cpu": [
+				"x64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"sunos"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/win32-arm64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/win32-arm64/-/win32-arm64-0.25.0.tgz",
+			"integrity": "sha512-ZUAc2YK6JW89xTbXvftxdnYy3m4iHIkDtK3CLce8wg8M2L+YZhIvO1DKpxrd0Yr59AeNNkTiic9YLf6FTtXWMw==",
+			"cpu": [
+				"arm64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"win32"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/win32-ia32": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/win32-ia32/-/win32-ia32-0.25.0.tgz",
+			"integrity": "sha512-eSNxISBu8XweVEWG31/JzjkIGbGIJN/TrRoiSVZwZ6pkC6VX4Im/WV2cz559/TXLcYbcrDN8JtKgd9DJVIo8GA==",
+			"cpu": [
+				"ia32"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"win32"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
+		"node_modules/@esbuild/win32-x64": {
+			"version": "0.25.0",
+			"resolved": "https://registry.npmjs.org/@esbuild/win32-x64/-/win32-x64-0.25.0.tgz",
+			"integrity": "sha512-ZENoHJBxA20C2zFzh6AI4fT6RraMzjYw4xKWemRTRmRVtN9c5DcH9r/f2ihEkMjOW5eGgrwCslG/+Y/3bL+DHQ==",
+			"cpu": [
+				"x64"
+			],
+			"dev": true,
+			"license": "MIT",
+			"optional": true,
+			"os": [
+				"win32"
+			],
+			"engines": {
+				"node": ">=18"
+			}
+		},
 		"node_modules/@eslint-community/eslint-utils": {
 			"version": "4.4.0",
 			"resolved": "https://registry.npmjs.org/@eslint-community/eslint-utils/-/eslint-utils-4.4.0.tgz",
@@ -3355,6 +3789,27 @@
 			"resolved": "https://registry.npmjs.org/@firebase/webchannel-wrapper/-/webchannel-wrapper-1.0.3.tgz",
 			"integrity": "sha512-2xCRM9q9FlzGZCdgDMJwc0gyUkWFtkosy7Xxr6sFgQwn+wMNIWd7xIvYNauU1r64B5L5rsGKy/n9TKJ0aAFeqQ=="
 		},
+		"node_modules/@google-cloud/vertexai": {
+			"version": "1.9.3",
+			"resolved": "https://registry.npmjs.org/@google-cloud/vertexai/-/vertexai-1.9.3.tgz",
+			"integrity": "sha512-35o5tIEMLW3JeFJOaaMNR2e5sq+6rpnhrF97PuAxeOm0GlqVTESKhkGj7a5B5mmJSSSU3hUfIhcQCRRsw4Ipzg==",
+			"license": "Apache-2.0",
+			"dependencies": {
+				"google-auth-library": "^9.1.0"
+			},
+			"engines": {
+				"node": ">=18.0.0"
+			}
+		},
+		"node_modules/@google/generative-ai": {
+			"version": "0.18.0",
+			"resolved": "https://registry.npmjs.org/@google/generative-ai/-/generative-ai-0.18.0.tgz",
+			"integrity": "sha512-AhaIWSpk2tuhYHrBhUqC0xrWWznmYEja1/TRDIb+5kruBU5kUzMlFsXCQNO9PzyTZ4clUJ3CX/Rvy+Xm9x+w3g==",
+			"license": "Apache-2.0",
+			"engines": {
+				"node": ">=18.0.0"
+			}
+		},
 		"node_modules/@grpc/grpc-js": {
 			"version": "1.9.15",
 			"resolved": "https://registry.npmjs.org/@grpc/grpc-js/-/grpc-js-1.9.15.tgz",
@@ -10610,6 +11065,15 @@
 				"url": "https://github.com/sponsors/ljharb"
 			}
 		},
+		"node_modules/ollama": {
+			"version": "0.5.13",
+			"resolved": "https://registry.npmjs.org/ollama/-/ollama-0.5.13.tgz",
+			"integrity": "sha512-qK3eE2GjMYjCiTknEJfAHjbUzUqgVtf9qtzjxWrkwBZgBG7kOB6Z4+Ov4fbvDjmKKHv+rpuTsWFg4jZvVjNBtQ==",
+			"license": "MIT",
+			"dependencies": {
+				"whatwg-fetch": "^3.6.20"
+			}
+		},
 		"node_modules/once": {
 			"version": "1.4.0",
 			"resolved": "https://registry.npmjs.org/once/-/once-1.4.0.tgz",
@@ -10651,9 +11115,9 @@
 			}
 		},
 		"node_modules/open-graph-scraper/node_modules/chardet": {
-			"version": "2.0.0",
-			"resolved": "https://registry.npmjs.org/chardet/-/chardet-2.0.0.tgz",
-			"integrity": "sha512-xVgPpulCooDjY6zH4m9YW3jbkaBe3FKIAvF5sj5t7aBNsVl2ljIE+xwJ4iNgiDZHFQvNIpjdKdVOQvvk5ZfxbQ==",
+			"version": "2.1.0",
+			"resolved": "https://registry.npmjs.org/chardet/-/chardet-2.1.0.tgz",
+			"integrity": "sha512-bNFETTG/pM5ryzQ9Ad0lJOTa6HWD/YsScAR3EnCPZRPlQh77JocYktSHOUHelyhm8IARL+o4c4F1bP5KVOjiRA==",
 			"license": "MIT"
 		},
 		"node_modules/openai": {
@@ -12960,6 +13424,12 @@
 				"node": ">=18"
 			}
 		},
+		"node_modules/whatwg-fetch": {
+			"version": "3.6.20",
+			"resolved": "https://registry.npmjs.org/whatwg-fetch/-/whatwg-fetch-3.6.20.tgz",
+			"integrity": "sha512-EqhiFU6daOA8kpjOWTL0olhVOF3i7OrFzSYiGsEMB8GcXS+RrzauAERX65xMeNWVqxA6HXH2m69Z9LaKKdisfg==",
+			"license": "MIT"
+		},
 		"node_modules/whatwg-mimetype": {
 			"version": "4.0.0",
 			"resolved": "https://registry.npmjs.org/whatwg-mimetype/-/whatwg-mimetype-4.0.0.tgz",
@@ -2,7 +2,7 @@
 	"name": "claude-dev",
 	"displayName": "Cline",
 	"description": "Autonomous coding agent right in your IDE, capable of creating/editing files, running commands, using the browser, and more with your permission every step of the way.",
-	"version": "3.4.9",
+	"version": "3.5.1",
 	"icon": "assets/icons/icon.png",
 	"galleryBanner": {
 		"color": "#617A91",
@@ -226,7 +226,7 @@
 		"watch-tests": "tsc -p . -w --outDir out",
 		"pretest": "npm run compile-tests && npm run compile && npm run lint",
 		"check-types": "tsc --noEmit",
-		"lint": "eslint src --ext ts",
+		"lint": "eslint src --ext ts && eslint webview-ui/src --ext ts",
 		"format": "prettier . --check",
 		"format:fix": "prettier . --write",
 		"test": "vscode-test",
@@ -265,6 +265,7 @@
 		"@anthropic-ai/bedrock-sdk": "^0.12.4",
 		"@anthropic-ai/sdk": "^0.37.0",
 		"@anthropic-ai/vertex-sdk": "^0.6.4",
+		"@google-cloud/vertexai": "^1.9.3",
 		"@google/generative-ai": "^0.18.0",
 		"@mistralai/mistralai": "^1.5.0",
 		"@modelcontextprotocol/sdk": "^1.0.1",
@@ -289,6 +290,7 @@
 		"isbinaryfile": "^5.0.2",
 		"mammoth": "^1.8.0",
 		"monaco-vscode-textmate-theme-converter": "^0.1.7",
+		"ollama": "^0.5.13",
 		"open-graph-scraper": "^6.9.0",
 		"openai": "^4.83.0",
 		"os-name": "^6.0.0",
@@ -17,6 +17,7 @@ import { QwenHandler } from "./providers/qwen"
 import { MistralHandler } from "./providers/mistral"
 import { VsCodeLmHandler } from "./providers/vscode-lm"
 import { LiteLlmHandler } from "./providers/litellm"
+import { AskSageHandler } from "./providers/asksage"
 import { XAIHandler } from "./providers/xai"

 export interface ApiHandler {
@@ -63,6 +64,8 @@ export function buildApiHandler(configuration: ApiConfiguration): ApiHandler {
 			return new VsCodeLmHandler(options)
 		case "litellm":
 			return new LiteLlmHandler(options)
+		case "asksage":
+			return new AskSageHandler(options)
 		case "xai":
 			return new XAIHandler(options)
 		default:
@@ -22,6 +22,10 @@ export class AnthropicHandler implements ApiHandler {
 		const model = this.getModel()
 		let stream: AnthropicStream<Anthropic.RawMessageStreamEvent>
 		const modelId = model.id
+
+		let budget_tokens = this.options.thinkingBudgetTokens || 0
+		const reasoningOn = modelId.includes("3-7") && budget_tokens !== 0 ? true : false
+
 		switch (modelId) {
 			// 'latest' alias does not support cache_control
 			case "claude-3-7-sonnet-20250219":
@@ -41,8 +45,11 @@ export class AnthropicHandler implements ApiHandler {
 				stream = await this.client.messages.create(
 					{
 						model: modelId,
+						thinking: reasoningOn ? { type: "enabled", budget_tokens: budget_tokens } : undefined,
 						max_tokens: model.info.maxTokens || 8192,
-						temperature: 0,
+						// "Thinking isn’t compatible with temperature, top_p, or top_k modifications as well as forced tool use."
+						// (https://docs.anthropic.com/en/docs/build-with-claude/extended-thinking#important-considerations-when-using-extended-thinking)
+						temperature: reasoningOn ? undefined : 0,
 						system: [
 							{
 								text: systemPrompt,
@@ -148,6 +155,20 @@ export class AnthropicHandler implements ApiHandler {
 					break
 				case "content_block_start":
 					switch (chunk.content_block.type) {
+						case "thinking":
+							yield {
+								type: "reasoning",
+								reasoning: chunk.content_block.thinking || "",
+							}
+							break
+						case "redacted_thinking":
+							// Handle redacted thinking blocks - we still mark it as reasoning
+							// but note that the content is encrypted
+							yield {
+								type: "reasoning",
+								reasoning: "[Redacted thinking block]",
+							}
+							break
 						case "text":
 							// we may receive multiple text blocks, in which case just insert a line break between them
 							if (chunk.index > 0) {
@@ -165,12 +186,22 @@ export class AnthropicHandler implements ApiHandler {
 					break
 				case "content_block_delta":
 					switch (chunk.delta.type) {
+						case "thinking_delta":
+							yield {
+								type: "reasoning",
+								reasoning: chunk.delta.thinking,
+							}
+							break
 						case "text_delta":
 							yield {
 								type: "text",
 								text: chunk.delta.text,
 							}
 							break
+						case "signature_delta":
+							// We don't need to do anything with the signature in the client
+							// It's used when sending the thinking block back to the API
+							break
 					}
 					break
 				case "content_block_stop":
@@ -0,0 +1,115 @@
+import { Anthropic } from "@anthropic-ai/sdk"
+import { ApiHandler } from ".."
+import {
+	ApiHandlerOptions,
+	ModelInfo,
+	AskSageModelId,
+	askSageModels,
+	askSageDefaultModelId,
+	askSageDefaultURL,
+} from "../../shared/api"
+import { ApiStream } from "../transform/stream"
+
+type AskSageRequest = {
+	system_prompt: string
+	message: {
+		user: "gpt" | "me"
+		message: string
+	}[]
+	model: string
+	dataset: "none"
+}
+
+type AskSageResponse = {
+	uuid: string
+	status: number
+	// Response status
+	response: string
+	// Generated response message
+	message: string
+}
+
+export class AskSageHandler implements ApiHandler {
+	private options: ApiHandlerOptions
+	private apiUrl: string
+	private apiKey: string
+
+	constructor(options: ApiHandlerOptions) {
+		console.log("init api url", options.asksageApiUrl, askSageDefaultURL)
+		this.options = options
+		this.apiKey = options.asksageApiKey || ""
+		this.apiUrl = options.asksageApiUrl || askSageDefaultURL
+
+		if (!this.apiKey) {
+			throw new Error("AskSage API key is required")
+		}
+	}
+
+	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+		try {
+			const model = this.getModel()
+
+			// Transform messages into AskSageRequest format
+			const formattedMessages = messages.map((msg) => {
+				const content = Array.isArray(msg.content)
+					? msg.content.map((block) => ("text" in block ? block.text : "")).join("")
+					: msg.content
+
+				return {
+					user: msg.role === "assistant" ? ("gpt" as const) : ("me" as const),
+					message: content,
+				}
+			})
+
+			const request: AskSageRequest = {
+				system_prompt: systemPrompt,
+				message: formattedMessages,
+				model: model.id,
+				dataset: "none",
+			}
+
+			// Make request to AskSage API
+			const response = await fetch(`${this.apiUrl}/query`, {
+				method: "POST",
+				headers: {
+					"Content-Type": "application/json",
+					"x-access-tokens": this.apiKey,
+				},
+				body: JSON.stringify(request),
+			})
+
+			if (!response.ok) {
+				const error = await response.text()
+				throw new Error(`AskSage API error: ${error}`)
+			}
+
+			const result = (await response.json()) as AskSageResponse
+
+			if (!result.message) {
+				throw new Error("No content in AskSage response")
+			}
+
+			// Return entire response as a single chunk since streaming is not supported
+			yield {
+				type: "text",
+				text: result.message,
+			}
+		} catch (error) {
+			if (error instanceof Error) {
+				throw new Error(`AskSage request failed: ${error.message}`)
+			}
+		}
+	}
+
+	getModel(): { id: string; info: ModelInfo } {
+		const modelId = this.options.apiModelId
+		if (modelId && modelId in askSageModels) {
+			const id = modelId as AskSageModelId
+			return { id, info: askSageModels[id] }
+		}
+		return {
+			id: askSageDefaultModelId,
+			info: askSageModels[askSageDefaultModelId],
+		}
+	}
+}
@@ -1,5 +1,6 @@
 import AnthropicBedrock from "@anthropic-ai/bedrock-sdk"
 import { Anthropic } from "@anthropic-ai/sdk"
+import { withRetry } from "../retry"
 import { ApiHandler } from "../"
 import { ApiHandlerOptions, bedrockDefaultModelId, BedrockModelId, bedrockModels, ModelInfo } from "../../shared/api"
 import { ApiStream } from "../transform/stream"
@@ -13,22 +14,70 @@ export class AwsBedrockHandler implements ApiHandler {
 		this.options = options
 	}

+	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		// cross region inference requires prefixing the model id with the region
 		let modelId = await this.getModelId()

-		// create anthropic client, using sessions created or renewed after this handler's
+		let budget_tokens = this.options.thinkingBudgetTokens || 0
+		const reasoningOn = modelId.includes("3-7") && budget_tokens !== 0 ? true : false
+
+		// Get model info and message indices for caching
+		const model = this.getModel()
+		const userMsgIndices = messages.reduce((acc, msg, index) => (msg.role === "user" ? [...acc, index] : acc), [] as number[])
+		const lastUserMsgIndex = userMsgIndices[userMsgIndices.length - 1] ?? -1
+		const secondLastMsgUserIndex = userMsgIndices[userMsgIndices.length - 2] ?? -1
+
+		// Create anthropic client, using sessions created or renewed after this handler's
 		// initialization, and allowing for session renewal if necessary as well
-		let client = await this.getClient()
+		const client = await this.getClient()

 		const stream = await client.messages.create({
 			model: modelId,
-			max_tokens: this.getModel().info.maxTokens || 8192,
-			temperature: 0,
-			system: systemPrompt,
-			messages,
+			max_tokens: model.info.maxTokens || 8192,
+			thinking: reasoningOn ? { type: "enabled", budget_tokens: budget_tokens } : undefined,
+			temperature: reasoningOn ? undefined : 0,
+			system: [
+				{
+					text: systemPrompt,
+					type: "text",
+					...(this.options.awsBedrockUsePromptCache === true && {
+						cache_control: { type: "ephemeral" },
+					}),
+				},
+			],
+			messages: messages.map((message, index) => {
+				if (index === lastUserMsgIndex || index === secondLastMsgUserIndex) {
+					return {
+						...message,
+						content:
+							typeof message.content === "string"
+								? [
+										{
+											type: "text",
+											text: message.content,
+											...(this.options.awsBedrockUsePromptCache === true && {
+												cache_control: { type: "ephemeral" },
+											}),
+										},
+									]
+								: message.content.map((content, contentIndex) =>
+										contentIndex === message.content.length - 1
+											? {
+													...content,
+													...(this.options.awsBedrockUsePromptCache === true && {
+														cache_control: { type: "ephemeral" },
+													}),
+												}
+											: content,
+									),
+					}
+				}
+				return message
+			}),
 			stream: true,
 		})
+
 		for await (const chunk of stream) {
 			switch (chunk.type) {
 				case "message_start":
@@ -37,6 +86,8 @@ export class AwsBedrockHandler implements ApiHandler {
 						type: "usage",
 						inputTokens: usage.input_tokens || 0,
 						outputTokens: usage.output_tokens || 0,
+						cacheWriteTokens: usage.cache_creation_input_tokens || undefined,
+						cacheReadTokens: usage.cache_read_input_tokens || undefined,
 					}
 					break
 				case "message_delta":
@@ -46,7 +97,6 @@ export class AwsBedrockHandler implements ApiHandler {
 						outputTokens: chunk.usage.output_tokens || 0,
 					}
 					break
-
 				case "content_block_start":
 					switch (chunk.content_block.type) {
 						case "text":
@@ -118,7 +168,7 @@ export class AwsBedrockHandler implements ApiHandler {
 		})
 	}

-	private async getModelId(): Promise<string> {
+	async getModelId(): Promise<string> {
 		if (this.options.awsUseCrossRegionInference) {
 			let regionPrefix = (this.options.awsRegion || "").slice(0, 3)
 			switch (regionPrefix) {
@@ -126,11 +176,11 @@ export class AwsBedrockHandler implements ApiHandler {
 					return `us.${this.getModel().id}`
 				case "eu-":
 					return `eu.${this.getModel().id}`
-					break
+				case "ap-":
+					return `apac.${this.getModel().id}`
 				default:
 					// cross region inference is not supported in this region, falling back to default model
 					return this.getModel().id
-					break
 			}
 		}
 		return this.getModel().id
@@ -147,7 +197,7 @@ export class AwsBedrockHandler implements ApiHandler {
 		}
 	}

-	private static async setEnv(key: string, value: string | undefined) {
+	private static setEnv(key: string, value: string | undefined) {
 		if (key !== "" && value !== undefined) {
 			process.env[key] = value
 		}
@@ -3,6 +3,7 @@ import OpenAI from "openai"
 import { withRetry } from "../retry"
 import { ApiHandler } from "../"
 import { ApiHandlerOptions, DeepSeekModelId, ModelInfo, deepSeekDefaultModelId, deepSeekModels } from "../../shared/api"
+import { calculateApiCostOpenAI } from "../../utils/cost"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { ApiStream } from "../transform/stream"
 import { convertToR1Format } from "../transform/r1-format"
@@ -19,6 +20,37 @@ export class DeepSeekHandler implements ApiHandler {
 		})
 	}

+	private async *yieldUsage(info: ModelInfo, usage: OpenAI.Completions.CompletionUsage | undefined): ApiStream {
+		// Deepseek reports total input AND cache reads/writes,
+		// see context caching: https://api-docs.deepseek.com/guides/kv_cache)
+		// where the input tokens is the sum of the cache hits/misses, just like OpenAI.
+		// This affects:
+		// 1) context management truncation algorithm, and
+		// 2) cost calculation
+
+		// Deepseek usage includes extra fields.
+		// Safely cast the prompt token details section to the appropriate structure.
+		interface DeepSeekUsage extends OpenAI.CompletionUsage {
+			prompt_cache_hit_tokens?: number
+			prompt_cache_miss_tokens?: number
+		}
+		const deepUsage = usage as DeepSeekUsage
+
+		const inputTokens = deepUsage?.prompt_tokens || 0
+		const outputTokens = deepUsage?.completion_tokens || 0
+		const cacheReadTokens = deepUsage?.prompt_cache_hit_tokens || 0
+		const cacheWriteTokens = deepUsage?.prompt_cache_miss_tokens || 0
+		const totalCost = calculateApiCostOpenAI(info, inputTokens, outputTokens, cacheWriteTokens, cacheReadTokens)
+		yield {
+			type: "usage",
+			inputTokens: inputTokens,
+			outputTokens: outputTokens,
+			cacheWriteTokens: cacheWriteTokens,
+			cacheReadTokens: cacheReadTokens,
+			totalCost: totalCost,
+		}
+	}
+
 	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		const model = this.getModel()
@@ -61,15 +93,7 @@ export class DeepSeekHandler implements ApiHandler {
 			}

 			if (chunk.usage) {
-				yield {
-					type: "usage",
-					inputTokens: chunk.usage.prompt_tokens || 0, // (deepseek reports total input AND cache reads/writes, see context caching: https://api-docs.deepseek.com/guides/kv_cache) where the input tokens is the sum of the cache hits/misses, while anthropic reports them as separate tokens. This is important to know for 1) context management truncation algorithm, and 2) cost calculation (NOTE: we report both input and cache stats but for now set input price to 0 since all the cost calculation will be done using cache hits/misses)
-					outputTokens: chunk.usage.completion_tokens || 0,
-					// @ts-ignore-next-line
-					cacheReadTokens: chunk.usage.prompt_cache_hit_tokens || 0,
-					// @ts-ignore-next-line
-					cacheWriteTokens: chunk.usage.prompt_cache_miss_tokens || 0,
-				}
+				yield* this.yieldUsage(model.info, chunk.usage)
 			}
 		}
 	}
@@ -1,40 +1,35 @@
 import { Anthropic } from "@anthropic-ai/sdk"
-import OpenAI from "openai"
+import { Message, Ollama } from "ollama"
 import { ApiHandler } from "../"
 import { ApiHandlerOptions, ModelInfo, openAiModelInfoSaneDefaults } from "../../shared/api"
-import { convertToOpenAiMessages } from "../transform/openai-format"
+import { convertToOllamaMessages } from "../transform/ollama-format"
 import { ApiStream } from "../transform/stream"

 export class OllamaHandler implements ApiHandler {
 	private options: ApiHandlerOptions
-	private client: OpenAI
+	private client: Ollama

 	constructor(options: ApiHandlerOptions) {
 		this.options = options
-		this.client = new OpenAI({
-			baseURL: (this.options.ollamaBaseUrl || "http://localhost:11434") + "/v1",
-			apiKey: "ollama",
-		})
+		this.client = new Ollama({ host: this.options.ollamaBaseUrl || "http://localhost:11434" })
 	}

 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
-		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [
-			{ role: "system", content: systemPrompt },
-			...convertToOpenAiMessages(messages),
-		]
+		const ollamaMessages: Message[] = [{ role: "system", content: systemPrompt }, ...convertToOllamaMessages(messages)]

-		const stream = await this.client.chat.completions.create({
+		const stream = await this.client.chat({
 			model: this.getModel().id,
-			messages: openAiMessages,
-			temperature: 0,
+			messages: ollamaMessages,
 			stream: true,
+			options: {
+				num_ctx: Number(this.options.ollamaApiOptionsCtxNum) || 32768,
+			},
 		})
 		for await (const chunk of stream) {
-			const delta = chunk.choices[0]?.delta
-			if (delta?.content) {
+			if (typeof chunk.message.content === "string") {
 				yield {
 					type: "text",
-					text: delta.content,
+					text: chunk.message.content,
 				}
 			}
 		}
@@ -10,6 +10,7 @@ import {
 	openAiNativeModels,
 } from "../../shared/api"
 import { convertToOpenAiMessages } from "../transform/openai-format"
+import { calculateApiCostOpenAI } from "../../utils/cost"
 import { ApiStream } from "../transform/stream"
 import { ChatCompletionReasoningEffort } from "openai/resources/chat/completions.mjs"

@@ -24,31 +25,47 @@ export class OpenAiNativeHandler implements ApiHandler {
 		})
 	}

+	private async *yieldUsage(info: ModelInfo, usage: OpenAI.Completions.CompletionUsage | undefined): ApiStream {
+		const inputTokens = usage?.prompt_tokens || 0
+		const outputTokens = usage?.completion_tokens || 0
+		const cacheReadTokens = usage?.prompt_tokens_details?.cached_tokens || 0
+		const cacheWriteTokens = 0
+		const totalCost = calculateApiCostOpenAI(info, inputTokens, outputTokens, cacheWriteTokens, cacheReadTokens)
+		yield {
+			type: "usage",
+			inputTokens: inputTokens,
+			outputTokens: outputTokens,
+			cacheWriteTokens: cacheWriteTokens,
+			cacheReadTokens: cacheReadTokens,
+			totalCost: totalCost,
+		}
+	}
+
 	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
-		switch (this.getModel().id) {
+		const model = this.getModel()
+
+		switch (model.id) {
 			case "o1":
 			case "o1-preview":
 			case "o1-mini": {
 				// o1 doesnt support streaming, non-1 temp, or system prompt
 				const response = await this.client.chat.completions.create({
-					model: this.getModel().id,
+					model: model.id,
 					messages: [{ role: "user", content: systemPrompt }, ...convertToOpenAiMessages(messages)],
 				})
 				yield {
 					type: "text",
 					text: response.choices[0]?.message.content || "",
 				}
-				yield {
-					type: "usage",
-					inputTokens: response.usage?.prompt_tokens || 0,
-					outputTokens: response.usage?.completion_tokens || 0,
-				}
+
+				yield* this.yieldUsage(model.info, response.usage)
+
 				break
 			}
 			case "o3-mini": {
 				const stream = await this.client.chat.completions.create({
-					model: this.getModel().id,
+					model: model.id,
 					messages: [{ role: "developer", content: systemPrompt }, ...convertToOpenAiMessages(messages)],
 					stream: true,
 					stream_options: { include_usage: true },
@@ -63,18 +80,15 @@ export class OpenAiNativeHandler implements ApiHandler {
 						}
 					}
 					if (chunk.usage) {
-						yield {
-							type: "usage",
-							inputTokens: chunk.usage.prompt_tokens || 0,
-							outputTokens: chunk.usage.completion_tokens || 0,
-						}
+						// Only last chunk contains usage
+						yield* this.yieldUsage(model.info, chunk.usage)
 					}
 				}
 				break
 			}
 			default: {
 				const stream = await this.client.chat.completions.create({
-					model: this.getModel().id,
+					model: model.id,
 					// max_completion_tokens: this.getModel().info.maxTokens,
 					temperature: 0,
 					messages: [{ role: "system", content: systemPrompt }, ...convertToOpenAiMessages(messages)],
@@ -90,14 +104,9 @@ export class OpenAiNativeHandler implements ApiHandler {
 							text: delta.content,
 						}
 					}
-
-					// contains a null value except for the last chunk which contains the token usage statistics for the entire request
 					if (chunk.usage) {
-						yield {
-							type: "usage",
-							inputTokens: chunk.usage.prompt_tokens || 0,
-							outputTokens: chunk.usage.completion_tokens || 0,
-						}
+						// Only last chunk contains usage
+						yield* this.yieldUsage(model.info, chunk.usage)
 					}
 				}
 			}
@@ -6,6 +6,7 @@ import { ApiHandler } from "../index"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { ApiStream } from "../transform/stream"
 import { convertToR1Format } from "../transform/r1-format"
+import { ChatCompletionReasoningEffort } from "openai/resources/chat/completions.mjs"

 export class OpenAiHandler implements ApiHandler {
 	private options: ApiHandlerOptions
@@ -28,110 +29,60 @@ export class OpenAiHandler implements ApiHandler {
 		}
 	}

-	private async diagnoseRequestProblem(
-		modelId: string,
-		messages: OpenAI.Chat.ChatCompletionMessageParam[],
-		apiKey: string,
-		baseURL: string,
-	) {
-		const url = `${baseURL}/chat/completions`
-
-		try {
-			const response = await fetch(url, {
-				method: "POST",
-				headers: {
-					"Content-Type": "application/json",
-					Authorization: `Bearer ${apiKey}`,
-				},
-				body: JSON.stringify({
-					model: modelId,
-					messages: messages,
-					temperature: 0,
-					stream: true,
-				}),
-			})
-
-			if (!response.ok) {
-				return `HTTP error! status: ${response.status}, statusText: ${response.statusText}`
-			}
-
-			const responseData = await response.json()
-			return responseData
-		} catch (error) {
-			return error instanceof Error ? error.message : String(error)
-		}
-	}
-
-	private async *handleChunk(chunk: OpenAI.Chat.Completions.ChatCompletionChunk): ApiStream {
-		const delta = chunk.choices[0]?.delta
-		if (delta?.content) {
-			yield {
-				type: "text",
-				text: delta.content,
-			}
-		}
-
-		if (delta && "reasoning_content" in delta && delta.reasoning_content) {
-			yield {
-				type: "reasoning",
-				reasoning: (delta.reasoning_content as string | undefined) || "",
-			}
-		}
-
-		if (chunk.usage) {
-			yield {
-				type: "usage",
-				inputTokens: chunk.usage.prompt_tokens || 0,
-				outputTokens: chunk.usage.completion_tokens || 0,
-			}
-		}
-	}
-
 	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		const modelId = this.options.openAiModelId ?? ""
 		const isDeepseekReasoner = modelId.includes("deepseek-reasoner")
+		const isO3Mini = modelId.includes("o3-mini")

 		let openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [
 			{ role: "system", content: systemPrompt },
 			...convertToOpenAiMessages(messages),
 		]
+		let temperature: number | undefined = 0
+		let reasoningEffort: ChatCompletionReasoningEffort | undefined = undefined

 		if (isDeepseekReasoner) {
 			openAiMessages = convertToR1Format([{ role: "user", content: systemPrompt }, ...messages])
 		}

+		if (isO3Mini) {
+			openAiMessages = [{ role: "developer", content: systemPrompt }, ...convertToOpenAiMessages(messages)]
+			temperature = undefined // does not support temperature
+			reasoningEffort = (this.options.o3MiniReasoningEffort as ChatCompletionReasoningEffort) || "medium"
+		}
+
 		const stream = await this.client.chat.completions.create({
 			model: modelId,
 			messages: openAiMessages,
-			temperature: 0,
+			temperature,
+			reasoning_effort: reasoningEffort,
 			stream: true,
 			stream_options: { include_usage: true },
 		})
+		for await (const chunk of stream) {
+			const delta = chunk.choices[0]?.delta
+			if (delta?.content) {
+				yield {
+					type: "text",
+					text: delta.content,
+				}
+			}

-		const [validationStream, contentStream] = stream.tee()
+			if (delta && "reasoning_content" in delta && delta.reasoning_content) {
+				yield {
+					type: "reasoning",
+					reasoning: (delta.reasoning_content as string | undefined) || "",
+				}
+			}

-		// Check the first chunk to detect potential stream issues early
-		// This helps to provide better error messages for cases like:
-		// https://github.com/cline/cline/issues/1662
-		// where the stream appears valid but contains no actual data
-		const firstChunk = await validationStream[Symbol.asyncIterator]().next()
-		if (firstChunk.done || !firstChunk.value) {
-			// Make an additional request to get detailed error information
-			// This gives us more context about what went wrong with the API call
-			const errorResponse = await this.diagnoseRequestProblem(
-				modelId,
-				openAiMessages,
-				this.client.apiKey,
-				this.client.baseURL,
-			)
-			throw new Error(`Stream empty. Error details: ${JSON.stringify(errorResponse)}`)
-		}
-
-		yield* this.handleChunk(firstChunk.value)
-
-		for await (const chunk of contentStream) {
-			yield* this.handleChunk(chunk)
+			if (chunk.usage) {
+				yield {
+					type: "usage",
+					inputTokens: chunk.usage.prompt_tokens || 0,
+					outputTokens: chunk.usage.completion_tokens || 0,
+				}
+			}
 		}
 	}

@@ -8,6 +8,7 @@ import { ApiHandlerOptions, ModelInfo, openRouterDefaultModelId, openRouterDefau
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { ApiStream } from "../transform/stream"
 import { convertToR1Format } from "../transform/r1-format"
+import { OpenRouterErrorResponse } from "./types"

 export class OpenRouterHandler implements ApiHandler {
 	private options: ApiHandlerOptions
@@ -40,6 +41,7 @@ export class OpenRouterHandler implements ApiHandler {
 		switch (model.id) {
 			case "anthropic/claude-3.7-sonnet":
 			case "anthropic/claude-3.7-sonnet:beta":
+			case "anthropic/claude-3.7-sonnet:thinking":
 			case "anthropic/claude-3-7-sonnet":
 			case "anthropic/claude-3-7-sonnet:beta":
 			case "anthropic/claude-3.5-sonnet":
@@ -95,6 +97,7 @@ export class OpenRouterHandler implements ApiHandler {
 		switch (model.id) {
 			case "anthropic/claude-3.7-sonnet":
 			case "anthropic/claude-3.7-sonnet:beta":
+			case "anthropic/claude-3.7-sonnet:thinking":
 			case "anthropic/claude-3-7-sonnet":
 			case "anthropic/claude-3-7-sonnet:beta":
 			case "anthropic/claude-3.5-sonnet":
@@ -109,7 +112,7 @@ export class OpenRouterHandler implements ApiHandler {
 				break
 		}

-		let temperature = 0
+		let temperature: number | undefined = 0
 		let topP: number | undefined = undefined
 		if (this.getModel().id.startsWith("deepseek/deepseek-r1") || this.getModel().id === "perplexity/sonar-reasoning") {
 			// Recommended values from DeepSeek
@@ -118,6 +121,22 @@ export class OpenRouterHandler implements ApiHandler {
 			openAiMessages = convertToR1Format([{ role: "user", content: systemPrompt }, ...messages])
 		}

+		let reasoning: { max_tokens: number } | undefined = undefined
+		switch (model.id) {
+			case "anthropic/claude-3.7-sonnet":
+			case "anthropic/claude-3.7-sonnet:beta":
+			case "anthropic/claude-3.7-sonnet:thinking":
+			case "anthropic/claude-3-7-sonnet":
+			case "anthropic/claude-3-7-sonnet:beta":
+				let budget_tokens = this.options.thinkingBudgetTokens || 0
+				const reasoningOn = budget_tokens !== 0 ? true : false
+				if (reasoningOn) {
+					temperature = undefined // extended thinking does not support non-1 temperature
+					reasoning = { max_tokens: budget_tokens }
+				}
+				break
+		}
+
 		// Removes messages in the middle when close to context window limit. Should not be applied to models that support prompt caching since it would continuously break the cache.
 		let shouldApplyMiddleOutTransform = !model.info.supportsPromptCache
 		// except for deepseek (which we set supportsPromptCache to true for), where because the context window is so small our truncation algo might miss and we should use openrouter's middle-out transform as a fallback to ensure we don't exceed the context window (FIXME: once we have a more robust token estimator we should not rely on this)
@@ -136,6 +155,7 @@ export class OpenRouterHandler implements ApiHandler {
 			transforms: shouldApplyMiddleOutTransform ? ["middle-out"] : undefined,
 			include_reasoning: true,
 			...(model.id === "openai/o3-mini" ? { reasoning_effort: this.options.o3MiniReasoningEffort || "medium" } : {}),
+			...(reasoning ? { reasoning } : {}),
 		})

 		let genId: string | undefined
@@ -143,9 +163,11 @@ export class OpenRouterHandler implements ApiHandler {
 		for await (const chunk of stream) {
 			// openrouter returns an error object instead of the openai sdk throwing an error
 			if ("error" in chunk) {
-				const error = chunk.error as { message?: string; code?: number }
+				const error = chunk.error as OpenRouterErrorResponse["error"]
 				console.error(`OpenRouter API Error: ${error?.code} - ${error?.message}`)
-				throw new Error(`OpenRouter API Error ${error?.code}: ${error?.message}`)
+				// Include metadata in the error message if available
+				const metadataStr = error.metadata ? `\nMetadata: ${JSON.stringify(error.metadata, null, 2)}` : ""
+				throw new Error(`OpenRouter API Error ${error.code}: ${error.message}${metadataStr}`)
 			}

 			if (!genId && chunk.id) {
@@ -1,15 +1,8 @@
 import { Anthropic } from "@anthropic-ai/sdk"
 import OpenAI from "openai"
-import { withRetry } from "../retry"
-import { calculateApiCostOpenAI } from "../../utils/cost"
-import {
-	ApiHandlerOptions,
-	ModelInfo,
-	openAiModelInfoSaneDefaults,
-	requestyDefaultModelId,
-	requestyDefaultModelInfo,
-} from "../../shared/api"
+import { ApiHandlerOptions, ModelInfo, openAiModelInfoSaneDefaults } from "../../shared/api"
 import { ApiHandler } from "../index"
+import { withRetry } from "../retry"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { ApiStream } from "../transform/stream"

@@ -31,7 +24,7 @@ export class RequestyHandler implements ApiHandler {

 	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
-		const model = this.getModel()
+		const modelId = this.options.requestyModelId ?? ""

 		let openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [
 			{ role: "system", content: systemPrompt },
@@ -40,13 +33,12 @@ export class RequestyHandler implements ApiHandler {

 		// @ts-ignore-next-line
 		const stream = await this.client.chat.completions.create({
-			model: model.id,
-			max_tokens: model.info.maxTokens || undefined,
+			model: modelId,
 			messages: openAiMessages,
 			temperature: 0,
 			stream: true,
 			stream_options: { include_usage: true },
-			...(model.id === "openai/o3-mini" ? { reasoning_effort: this.options.o3MiniReasoningEffort || "medium" } : {}),
+			...(modelId === "openai/o3-mini" ? { reasoning_effort: this.options.o3MiniReasoningEffort || "medium" } : {}),
 		})

 		for await (const chunk of stream) {
@@ -77,24 +69,28 @@ export class RequestyHandler implements ApiHandler {

 			if (chunk.usage) {
 				const usage = chunk.usage as RequestyUsage
+				const inputTokens = usage.prompt_tokens || 0
+				const outputTokens = usage.completion_tokens || 0
+				const cacheWriteTokens = usage.prompt_tokens_details?.caching_tokens || undefined
+				const cacheReadTokens = usage.prompt_tokens_details?.cached_tokens || undefined
+				const totalCost = 0 // TODO: Replace with calculateApiCostOpenAI(model.info, inputTokens, outputTokens, cacheWriteTokens, cacheReadTokens)
+
 				yield {
 					type: "usage",
-					inputTokens: usage.prompt_tokens || 0,
-					outputTokens: usage.completion_tokens || 0,
-					cacheWriteTokens: usage.prompt_tokens_details?.caching_tokens || undefined,
-					cacheReadTokens: usage.prompt_tokens_details?.cached_tokens || undefined,
-					totalCost: usage.total_cost || undefined,
+					inputTokens: inputTokens,
+					outputTokens: outputTokens,
+					cacheWriteTokens: cacheWriteTokens,
+					cacheReadTokens: cacheReadTokens,
+					totalCost: totalCost,
 				}
 			}
 		}
 	}

 	getModel(): { id: string; info: ModelInfo } {
-		const modelId = this.options.requestyModelId
-		const modelInfo = this.options.requestyModelInfo
-		if (modelId && modelInfo) {
-			return { id: modelId, info: modelInfo }
+		return {
+			id: this.options.requestyModelId ?? "",
+			info: openAiModelInfoSaneDefaults,
 		}
-		return { id: requestyDefaultModelId, info: requestyDefaultModelInfo }
 	}
 }
@@ -0,0 +1,22 @@
+// For the following openrouter error type sources, see the docs here:
+// https://openrouter.ai/docs/api-reference/errors
+
+export type OpenRouterErrorResponse = {
+	error: {
+		message: string
+		code: number
+		metadata?: OpenRouterProviderErrorMetadata | OpenRouterModerationErrorMetadata | Record<string, unknown>
+	}
+}
+
+export type OpenRouterProviderErrorMetadata = {
+	provider_name: string // The name of the provider that encountered the error
+	raw: unknown // The raw error from the provider
+}
+
+export type OpenRouterModerationErrorMetadata = {
+	reasons: string[] // Why your input was flagged
+	flagged_input: string // The text segment that was flagged, limited to 100 characters. If the flagged input is longer than 100 characters, it will be truncated in the middle and replaced with ...
+	provider_name: string // The name of the provider that requested moderation
+	model_slug: string
+}
@@ -4,75 +4,246 @@ import { withRetry } from "../retry"
 import { ApiHandler } from "../"
 import { ApiHandlerOptions, ModelInfo, vertexDefaultModelId, VertexModelId, vertexModels } from "../../shared/api"
 import { ApiStream } from "../transform/stream"
+import { VertexAI } from "@google-cloud/vertexai"

 // https://docs.anthropic.com/en/api/claude-on-vertex-ai
 export class VertexHandler implements ApiHandler {
 	private options: ApiHandlerOptions
-	private client: AnthropicVertex
+	private clientAnthropic: AnthropicVertex
+	private clientVertex: VertexAI

 	constructor(options: ApiHandlerOptions) {
 		this.options = options
-		this.client = new AnthropicVertex({
+		this.clientAnthropic = new AnthropicVertex({
 			projectId: this.options.vertexProjectId,
 			// https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/use-claude#regions
 			region: this.options.vertexRegion,
 		})
+		this.clientVertex = new VertexAI({
+			project: this.options.vertexProjectId,
+			location: this.options.vertexRegion,
+		})
 	}

 	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
-		const stream = await this.client.messages.create({
-			model: this.getModel().id,
-			max_tokens: this.getModel().info.maxTokens || 8192,
-			temperature: 0,
-			system: systemPrompt,
-			messages,
-			stream: true,
-		})
-		for await (const chunk of stream) {
-			switch (chunk.type) {
-				case "message_start":
-					const usage = chunk.message.usage
-					yield {
-						type: "usage",
-						inputTokens: usage.input_tokens || 0,
-						outputTokens: usage.output_tokens || 0,
-					}
-					break
-				case "message_delta":
-					yield {
-						type: "usage",
-						inputTokens: 0,
-						outputTokens: chunk.usage.output_tokens || 0,
-					}
-					break
+		const model = this.getModel()
+		const modelId = model.id

-				case "content_block_start":
-					switch (chunk.content_block.type) {
-						case "text":
-							if (chunk.index > 0) {
+		if (modelId.includes("claude")) {
+			let budget_tokens = this.options.thinkingBudgetTokens || 0
+			const reasoningOn = modelId.includes("3-7") && budget_tokens !== 0 ? true : false
+
+			let stream
+			switch (modelId) {
+				case "claude-3-7-sonnet@20250219":
+				case "claude-3-5-sonnet-v2@20241022":
+				case "claude-3-5-sonnet@20240620":
+				case "claude-3-5-haiku@20241022":
+				case "claude-3-opus@20240229":
+				case "claude-3-haiku@20240307": {
+					// Find indices of user messages for cache control
+					const userMsgIndices = messages.reduce(
+						(acc, msg, index) => (msg.role === "user" ? [...acc, index] : acc),
+						[] as number[],
+					)
+					const lastUserMsgIndex = userMsgIndices[userMsgIndices.length - 1] ?? -1
+					const secondLastMsgUserIndex = userMsgIndices[userMsgIndices.length - 2] ?? -1
+
+					stream = await this.clientAnthropic.beta.messages.create(
+						{
+							model: modelId,
+							max_tokens: model.info.maxTokens || 8192,
+							thinking: reasoningOn ? { type: "enabled", budget_tokens: budget_tokens } : undefined,
+							temperature: reasoningOn ? undefined : 0,
+							system: [
+								{
+									text: systemPrompt,
+									type: "text",
+									cache_control: { type: "ephemeral" },
+								},
+							],
+							messages: messages.map((message, index) => {
+								if (index === lastUserMsgIndex || index === secondLastMsgUserIndex) {
+									return {
+										...message,
+										content:
+											typeof message.content === "string"
+												? [
+														{
+															type: "text",
+															text: message.content,
+															cache_control: {
+																type: "ephemeral",
+															},
+														},
+													]
+												: message.content.map((content, contentIndex) =>
+														contentIndex === message.content.length - 1
+															? {
+																	...content,
+																	cache_control: {
+																		type: "ephemeral",
+																	},
+																}
+															: content,
+													),
+									}
+								}
+								return {
+									...message,
+									content:
+										typeof message.content === "string"
+											? [
+													{
+														type: "text",
+														text: message.content,
+													},
+												]
+											: message.content,
+								}
+							}),
+							stream: true,
+						},
+						{
+							headers: {},
+						},
+					)
+					break
+				}
+				default: {
+					stream = await this.clientAnthropic.beta.messages.create({
+						model: modelId,
+						max_tokens: model.info.maxTokens || 8192,
+						temperature: 0,
+						system: [
+							{
+								text: systemPrompt,
+								type: "text",
+							},
+						],
+						messages: messages.map((message) => ({
+							...message,
+							content:
+								typeof message.content === "string"
+									? [
+											{
+												type: "text",
+												text: message.content,
+											},
+										]
+									: message.content,
+						})),
+						stream: true,
+					})
+					break
+				}
+			}
+			for await (const chunk of stream) {
+				switch (chunk.type) {
+					case "message_start":
+						const usage = chunk.message.usage
+						yield {
+							type: "usage",
+							inputTokens: usage.input_tokens || 0,
+							outputTokens: usage.output_tokens || 0,
+							cacheWriteTokens: usage.cache_creation_input_tokens || undefined,
+							cacheReadTokens: usage.cache_read_input_tokens || undefined,
+						}
+						break
+					case "message_delta":
+						yield {
+							type: "usage",
+							inputTokens: 0,
+							outputTokens: chunk.usage.output_tokens || 0,
+						}
+						break
+					case "message_stop":
+						break
+					case "content_block_start":
+						switch (chunk.content_block.type) {
+							case "text":
+								if (chunk.index > 0) {
+									yield {
+										type: "text",
+										text: "\n",
+									}
+								}
 								yield {
 									type: "text",
-									text: "\n",
+									text: chunk.content_block.text,
 								}
+								break
+						}
+						break
+					case "content_block_delta":
+						switch (chunk.delta.type) {
+							case "text_delta":
+								yield {
+									type: "text",
+									text: chunk.delta.text,
+								}
+								break
+						}
+						break
+					case "content_block_stop":
+						break
+				}
+			}
+		} else {
+			// gemini
+			const generativeModel = this.clientVertex.getGenerativeModel({
+				model: this.getModel().id,
+				systemInstruction: {
+					role: "system",
+					parts: [{ text: systemPrompt }],
+				},
+			})
+			const request = {
+				contents: [
+					{
+						role: "user",
+						parts: messages.map((m) => {
+							if (typeof m.content === "string") {
+								return { text: m.content }
+							} else if (Array.isArray(m.content)) {
+								return {
+									text: m.content
+										.map((block) => {
+											if (typeof block === "string") {
+												return block
+											} else if (block.type === "text") {
+												return block.text
+											} else {
+												console.log("Unsupported block type", block)
+												return ""
+											}
+										})
+										.join(" "),
+								}
+							} else {
+								return { text: "" }
 							}
+						}),
+					},
+				],
+			}
+			const streamingResult = await generativeModel.generateContentStream(request)
+			for await (const chunk of streamingResult.stream) {
+				// If usage data is available, yield it similarly:
+				// yield { type: "usage", inputTokens: 0, outputTokens: 0 }
+				// Otherwise, just yield text:
+				const candidates = chunk.candidates || []
+				for (const candidate of candidates) {
+					for (const part of candidate.content?.parts || []) {
+						if (part.text) {
 							yield {
 								type: "text",
-								text: chunk.content_block.text,
+								text: part.text,
 							}
-							break
+						}
 					}
-					break
-				case "content_block_delta":
-					switch (chunk.delta.type) {
-						case "text_delta":
-							yield {
-								type: "text",
-								text: chunk.delta.text,
-							}
-							break
-					}
-					break
+				}
 			}
 		}
 	}
@@ -1,7 +1,7 @@
 import { Anthropic } from "@anthropic-ai/sdk"
 import * as vscode from "vscode"
 import { ApiHandler, SingleCompletionHandler } from "../"
-import { calculateApiCost } from "../../utils/cost"
+import { calculateApiCostAnthropic } from "../../utils/cost"
 import { ApiStream } from "../transform/stream"
 import { convertToVsCodeLmMessages } from "../transform/vscode-lm-format"
 import { SELECTOR_SEPARATOR, stringifyVsCodeLmModelSelector } from "../../shared/vsCodeSelectorUtils"
@@ -525,7 +525,7 @@ export class VsCodeLmHandler implements ApiHandler, SingleCompletionHandler {
 				type: "usage",
 				inputTokens: totalInputTokens,
 				outputTokens: totalOutputTokens,
-				totalCost: calculateApiCost(this.getModel().info, totalInputTokens, totalOutputTokens),
+				totalCost: calculateApiCostAnthropic(this.getModel().info, totalInputTokens, totalOutputTokens),
 			}
 		} catch (error: unknown) {
 			this.ensureCleanState()
@@ -0,0 +1,109 @@
+import { Anthropic } from "@anthropic-ai/sdk"
+import { Message } from "ollama"
+
+export function convertToOllamaMessages(anthropicMessages: Anthropic.Messages.MessageParam[]): Message[] {
+	const ollamaMessages: Message[] = []
+
+	for (const anthropicMessage of anthropicMessages) {
+		if (typeof anthropicMessage.content === "string") {
+			ollamaMessages.push({
+				role: anthropicMessage.role,
+				content: anthropicMessage.content,
+			})
+		} else {
+			if (anthropicMessage.role === "user") {
+				const { nonToolMessages, toolMessages } = anthropicMessage.content.reduce<{
+					nonToolMessages: (Anthropic.TextBlockParam | Anthropic.ImageBlockParam)[]
+					toolMessages: Anthropic.ToolResultBlockParam[]
+				}>(
+					(acc, part) => {
+						if (part.type === "tool_result") {
+							acc.toolMessages.push(part)
+						} else if (part.type === "text" || part.type === "image") {
+							acc.nonToolMessages.push(part)
+						}
+						return acc
+					},
+					{ nonToolMessages: [], toolMessages: [] },
+				)
+
+				// Process tool result messages FIRST since they must follow the tool use messages
+				let toolResultImages: string[] = []
+				toolMessages.forEach((toolMessage) => {
+					// The Anthropic SDK allows tool results to be a string or an array of text and image blocks, enabling rich and structured content. In contrast, the Ollama SDK only supports tool results as a single string, so we map the Anthropic tool result parts into one concatenated string to maintain compatibility.
+					let content: string
+
+					if (typeof toolMessage.content === "string") {
+						content = toolMessage.content
+					} else {
+						content =
+							toolMessage.content
+								?.map((part) => {
+									if (part.type === "image") {
+										toolResultImages.push(`data:${part.source.media_type};base64,${part.source.data}`)
+										return "(see following user message for image)"
+									}
+									return part.text
+								})
+								.join("\n") ?? ""
+					}
+					ollamaMessages.push({
+						role: "user",
+						images: toolResultImages.length > 0 ? toolResultImages : undefined,
+						content: content,
+					})
+				})
+
+				// Process non-tool messages
+				if (nonToolMessages.length > 0) {
+					ollamaMessages.push({
+						role: "user",
+						content: nonToolMessages
+							.map((part) => {
+								if (part.type === "image") {
+									return `data:${part.source.media_type};base64,${part.source.data}`
+								}
+								return part.text
+							})
+							.join("\n"),
+					})
+				}
+			} else if (anthropicMessage.role === "assistant") {
+				const { nonToolMessages, toolMessages } = anthropicMessage.content.reduce<{
+					nonToolMessages: (Anthropic.TextBlockParam | Anthropic.ImageBlockParam)[]
+					toolMessages: Anthropic.ToolUseBlockParam[]
+				}>(
+					(acc, part) => {
+						if (part.type === "tool_use") {
+							acc.toolMessages.push(part)
+						} else if (part.type === "text" || part.type === "image") {
+							acc.nonToolMessages.push(part)
+						} // assistant cannot send tool_result messages
+						return acc
+					},
+					{ nonToolMessages: [], toolMessages: [] },
+				)
+
+				// Process non-tool messages
+				let content: string = ""
+				if (nonToolMessages.length > 0) {
+					content = nonToolMessages
+						.map((part) => {
+							if (part.type === "image") {
+								return "" // impossible as the assistant cannot send images
+							}
+							return part.text
+						})
+						.join("\n")
+				}
+
+				ollamaMessages.push({
+					role: "assistant",
+					content,
+				})
+			}
+		}
+	}
+
+	return ollamaMessages
+}
@@ -14,7 +14,7 @@ import { OpenRouterHandler } from "../api/providers/openrouter"
 import { ApiStream } from "../api/transform/stream"
 import CheckpointTracker from "../integrations/checkpoints/CheckpointTracker"
 import { DIFF_VIEW_URI_SCHEME, DiffViewProvider } from "../integrations/editor/DiffViewProvider"
-import { findToolName, formatContentBlockToMarkdown } from "../integrations/misc/export-markdown"
+import { formatContentBlockToMarkdown } from "../integrations/misc/export-markdown"
 import { extractTextFromFile } from "../integrations/misc/extract-text"
 import { showSystemNotification } from "../integrations/notifications"
 import { TerminalManager } from "../integrations/terminal/TerminalManager"
@@ -47,7 +47,7 @@ import {
 import { getApiMetrics } from "../shared/getApiMetrics"
 import { HistoryItem } from "../shared/HistoryItem"
 import { ClineAskResponse, ClineCheckpointRestore } from "../shared/WebviewMessage"
-import { calculateApiCost } from "../utils/cost"
+import { calculateApiCostAnthropic } from "../utils/cost"
 import { fileExistsAtPath } from "../utils/fs"
 import { arePathsEqual, getReadablePath } from "../utils/path"
 import { fixModelHtmlEscaping, removeInvalidChars } from "../utils/string"
@@ -75,7 +75,6 @@ export class Cline {
 	browserSession: BrowserSession
 	private didEditFile: boolean = false
 	customInstructions?: string
-	preferredLanguage?: LanguageKey
 	autoApprovalSettings: AutoApprovalSettings
 	private browserSettings: BrowserSettings
 	private chatSettings: ChatSettings
@@ -136,9 +135,6 @@ export class Cline {
 		this.browserSession = new BrowserSession(provider.context, browserSettings)
 		this.diffViewProvider = new DiffViewProvider(cwd)
 		this.customInstructions = customInstructions
-		this.preferredLanguage = getLanguageKey(
-			vscode.workspace.getConfiguration("cline").get<LanguageDisplay>("preferredLanguage"),
-		)
 		this.autoApprovalSettings = autoApprovalSettings
 		this.browserSettings = browserSettings
 		this.chatSettings = chatSettings
@@ -152,16 +148,6 @@ export class Cline {
 		} else {
 			throw new Error("Either historyItem or task/images must be provided")
 		}
-		// capture start of thread with the state at the beginning
-		telemetryService.capture({
-			event: "cline created",
-			properties: {
-				taskId: this.taskId,
-				isHistory: !!historyItem,
-				chatMode: this.chatSettings.mode,
-				hasImages: !!images,
-			},
-		})
 	}

 	updateBrowserSettings(browserSettings: BrowserSettings) {
@@ -299,10 +285,7 @@ export class Cline {
 			case "workspace":
 				if (!this.checkpointTracker) {
 					try {
-						this.checkpointTracker = await CheckpointTracker.create(
-							this.taskId,
-							this.providerRef.deref()?.context.globalStorageUri.fsPath,
-						)
+						this.checkpointTracker = await CheckpointTracker.create(this.taskId, this.providerRef.deref())
 						this.checkpointTrackerErrorMessage = undefined
 					} catch (error) {
 						const errorMessage = error instanceof Error ? error.message : "Unknown error"
@@ -415,10 +398,7 @@ export class Cline {
 		// TODO: handle if this is called from outside original workspace, in which case we need to show user error message we cant show diff outside of workspace?
 		if (!this.checkpointTracker) {
 			try {
-				this.checkpointTracker = await CheckpointTracker.create(
-					this.taskId,
-					this.providerRef.deref()?.context.globalStorageUri.fsPath,
-				)
+				this.checkpointTracker = await CheckpointTracker.create(this.taskId, this.providerRef.deref())
 				this.checkpointTrackerErrorMessage = undefined
 			} catch (error) {
 				const errorMessage = error instanceof Error ? error.message : "Unknown error"
@@ -522,10 +502,7 @@ export class Cline {

 		if (!this.checkpointTracker) {
 			try {
-				this.checkpointTracker = await CheckpointTracker.create(
-					this.taskId,
-					this.providerRef.deref()?.context.globalStorageUri.fsPath,
-				)
+				this.checkpointTracker = await CheckpointTracker.create(this.taskId, this.providerRef.deref())
 				this.checkpointTrackerErrorMessage = undefined
 			} catch (error) {
 				const errorMessage = error instanceof Error ? error.message : "Unknown error"
@@ -883,42 +860,7 @@ export class Cline {

 		// need to make sure that the api conversation history can be resumed by the api, even if it goes out of sync with cline messages

-		let existingApiConversationHistory: Anthropic.Messages.MessageParam[] = await this.getSavedApiConversationHistory()
-
-		// v2.0 xml tags refactor caveat: since we don't use tools anymore, we need to replace all tool use blocks with a text block since the API disallows conversations with tool uses and no tool schema
-		const conversationWithoutToolBlocks = existingApiConversationHistory.map((message) => {
-			if (Array.isArray(message.content)) {
-				const newContent = message.content.map((block) => {
-					if (block.type === "tool_use") {
-						// it's important we convert to the new tool schema format so the model doesn't get confused about how to invoke tools
-						const inputAsXml = Object.entries(block.input as Record<string, string>)
-							.map(([key, value]) => `<${key}>\n${value}\n</${key}>`)
-							.join("\n")
-						return {
-							type: "text",
-							text: `<${block.name}>\n${inputAsXml}\n</${block.name}>`,
-						} as Anthropic.Messages.TextBlockParam
-					} else if (block.type === "tool_result") {
-						// Convert block.content to text block array, removing images
-						const contentAsTextBlocks = Array.isArray(block.content)
-							? block.content.filter((item) => item.type === "text")
-							: [{ type: "text", text: block.content }]
-						const textContent = contentAsTextBlocks.map((item) => item.text).join("\n\n")
-						const toolName = findToolName(block.tool_use_id, existingApiConversationHistory)
-						return {
-							type: "text",
-							text: `[${toolName} Result]\n\n${textContent}`,
-						} as Anthropic.Messages.TextBlockParam
-					}
-					return block
-				})
-				return { ...message, content: newContent }
-			}
-			return message
-		})
-		existingApiConversationHistory = conversationWithoutToolBlocks
-
-		// FIXME: remove tool use blocks altogether
+		const existingApiConversationHistory: Anthropic.Messages.MessageParam[] = await this.getSavedApiConversationHistory()

 		// if the last message is an assistant message, we need to check if there's tool use since every tool use has to have a tool response
 		// if there's no tool use and only a text block, then we can just add a user message
@@ -1100,67 +1042,74 @@ export class Cline {
 	// Checkpoints

 	async saveCheckpoint(isAttemptCompletionMessage: boolean = false) {
-		const commitHash = await this.checkpointTracker?.commit() // silently fails for now
 		// Set isCheckpointCheckedOut to false for all checkpoint_created messages
 		this.clineMessages.forEach((message) => {
 			if (message.say === "checkpoint_created") {
 				message.isCheckpointCheckedOut = false
 			}
 		})
-		if (commitHash) {
-			if (!isAttemptCompletionMessage) {
-				// For non-attempt completion we just say checkpoints
-				await this.say("checkpoint_created", commitHash)
+
+		if (!isAttemptCompletionMessage) {
+			// For non-attempt completion we just say checkpoints
+			await this.say("checkpoint_created")
+			this.checkpointTracker?.commit().then(async (commitHash) => {
 				const lastCheckpointMessage = findLast(this.clineMessages, (m) => m.say === "checkpoint_created")
 				if (lastCheckpointMessage) {
 					lastCheckpointMessage.lastCheckpointHash = commitHash
 					await this.saveClineMessages()
 				}
-			} else {
-				// For attempt_completion, find the last completion_result message and set its checkpoint hash. This will be used to present the 'see new changes' button
-				const lastCompletionResultMessage = findLast(
-					this.clineMessages,
-					(m) => m.say === "completion_result" || m.ask === "completion_result",
-				)
-				if (lastCompletionResultMessage) {
-					lastCompletionResultMessage.lastCheckpointHash = commitHash
-					await this.saveClineMessages()
-				}
+			}) // silently fails for now
+
+			//
+		} else {
+			// attempt completion requires checkpoint to be sync so that we can present button after attempt_completion
+			const commitHash = await this.checkpointTracker?.commit()
+			// For attempt_completion, find the last completion_result message and set its checkpoint hash. This will be used to present the 'see new changes' button
+			const lastCompletionResultMessage = findLast(
+				this.clineMessages,
+				(m) => m.say === "completion_result" || m.ask === "completion_result",
+			)
+			if (lastCompletionResultMessage) {
+				lastCompletionResultMessage.lastCheckpointHash = commitHash
+				await this.saveClineMessages()
 			}
-
-			// Previously we checkpointed every message, but this is excessive and unnecessary.
-			// // Start from the end and work backwards until we find a tool use or another message with a hash
-			// for (let i = this.clineMessages.length - 1; i >= 0; i--) {
-			// 	const message = this.clineMessages[i]
-			// 	if (message.lastCheckpointHash) {
-			// 		// Found a message with a hash, so we can stop
-			// 		break
-			// 	}
-			// 	// Update this message with a hash
-			// 	message.lastCheckpointHash = commitHash
-
-			// 	// We only care about adding the hash to the last tool use (we don't want to add this hash to every prior message ie for tasks pre-checkpoint)
-			// 	const isToolUse =
-			// 		message.say === "tool" ||
-			// 		message.ask === "tool" ||
-			// 		message.say === "command" ||
-			// 		message.ask === "command" ||
-			// 		message.say === "completion_result" ||
-			// 		message.ask === "completion_result" ||
-			// 		message.ask === "followup" ||
-			// 		message.say === "use_mcp_server" ||
-			// 		message.ask === "use_mcp_server" ||
-			// 		message.say === "browser_action" ||
-			// 		message.say === "browser_action_launch" ||
-			// 		message.ask === "browser_action_launch"
-
-			// 	if (isToolUse) {
-			// 		break
-			// 	}
-			// }
-			// // Save the updated messages
-			// await this.saveClineMessages()
 		}
+
+		// if (commitHash) {
+
+		// Previously we checkpointed every message, but this is excessive and unnecessary.
+		// // Start from the end and work backwards until we find a tool use or another message with a hash
+		// for (let i = this.clineMessages.length - 1; i >= 0; i--) {
+		// 	const message = this.clineMessages[i]
+		// 	if (message.lastCheckpointHash) {
+		// 		// Found a message with a hash, so we can stop
+		// 		break
+		// 	}
+		// 	// Update this message with a hash
+		// 	message.lastCheckpointHash = commitHash
+
+		// 	// We only care about adding the hash to the last tool use (we don't want to add this hash to every prior message ie for tasks pre-checkpoint)
+		// 	const isToolUse =
+		// 		message.say === "tool" ||
+		// 		message.ask === "tool" ||
+		// 		message.say === "command" ||
+		// 		message.ask === "command" ||
+		// 		message.say === "completion_result" ||
+		// 		message.ask === "completion_result" ||
+		// 		message.ask === "followup" ||
+		// 		message.say === "use_mcp_server" ||
+		// 		message.ask === "use_mcp_server" ||
+		// 		message.say === "browser_action" ||
+		// 		message.say === "browser_action_launch" ||
+		// 		message.ask === "browser_action_launch"
+
+		// 	if (isToolUse) {
+		// 		break
+		// 	}
+		// }
+		// // Save the updated messages
+		// await this.saveClineMessages()
+		// }
 	}

 	// Tools
@@ -1292,9 +1241,12 @@ export class Cline {
 		let systemPrompt = await SYSTEM_PROMPT(cwd, supportsComputerUse, mcpHub, this.browserSettings)

 		let settingsCustomInstructions = this.customInstructions?.trim()
+		const preferredLanguage = getLanguageKey(
+			vscode.workspace.getConfiguration("cline").get<LanguageDisplay>("preferredLanguage"),
+		)
 		const preferredLanguageInstructions =
-			this.preferredLanguage && this.preferredLanguage !== DEFAULT_LANGUAGE_SETTINGS
-				? `# Preferred Language\n\nSpeak in ${this.preferredLanguage}.`
+			preferredLanguage && preferredLanguage !== DEFAULT_LANGUAGE_SETTINGS
+				? `# Preferred Language\n\nSpeak in ${preferredLanguage}.`
 				: ""
 		const clineRulesFilePath = path.resolve(cwd, GlobalFileNames.clineRules)
 		let clineRulesFileInstructions: string | undefined
@@ -1318,8 +1270,8 @@ export class Cline {
 		if (
 			settingsCustomInstructions ||
 			clineRulesFileInstructions ||
-			preferredLanguageInstructions ||
-			clineIgnoreInstructions
+			clineIgnoreInstructions ||
+			preferredLanguageInstructions
 		) {
 			// altering the system prompt mid-task will break the prompt cache, but in the grand scheme this will not change often so it's better to not pollute user messages with it the way we have to with <potentially relevant details>
 			systemPrompt += addUserInstructions(
@@ -2947,7 +2899,7 @@ export class Cline {
 		}

 		/*
-		Seeing out of bounds is fine, it means that the next too call is being built up and ready to add to assistantMessageContent to present.
+		Seeing out of bounds is fine, it means that the next too call is being built up and ready to add to assistantMessageContent to present. 
 		When you see the UI inactive during this, it means that a tool is breaking without presenting any UI. For example the write_to_file tool was breaking when relpath was undefined, and for invalid relpath it never presented UI.
 		*/
 		this.presentAssistantMessageLocked = false // this needs to be placed here, if not then calling this.presentAssistantMessage below would fail (sometimes) since it's locked
@@ -2997,7 +2949,7 @@ export class Cline {
 				"mistake_limit_reached",
 				this.api.getModel().id.includes("claude")
 					? `This may indicate a failure in his thought process or inability to use a tool properly, which can be mitigated with some user guidance (e.g. "Try breaking down the task into smaller steps").`
-					: "Cline uses complex prompts and iterative task execution that may be challenging for less capable models. For best results, it's recommended to use Claude 3.5 Sonnet for its advanced agentic coding capabilities.",
+					: "Cline uses complex prompts and iterative task execution that may be challenging for less capable models. For best results, it's recommended to use Claude 3.7 Sonnet for its advanced agentic coding capabilities.",
 			)
 			if (response === "messageResponse") {
 				userContent.push(
@@ -3054,10 +3006,7 @@ export class Cline {
 		// isNewTask &&
 		if (!this.checkpointTracker) {
 			try {
-				this.checkpointTracker = await CheckpointTracker.create(
-					this.taskId,
-					this.providerRef.deref()?.context.globalStorageUri.fsPath,
-				)
+				this.checkpointTracker = await CheckpointTracker.create(this.taskId, this.providerRef.deref())
 				this.checkpointTrackerErrorMessage = undefined
 			} catch (error) {
 				const errorMessage = error instanceof Error ? error.message : "Unknown error"
@@ -3113,7 +3062,13 @@ export class Cline {
 					cacheReads: cacheReadTokens,
 					cost:
 						totalCost ??
-						calculateApiCost(this.api.getModel().info, inputTokens, outputTokens, cacheWriteTokens, cacheReadTokens),
+						calculateApiCostAnthropic(
+							this.api.getModel().info,
+							inputTokens,
+							outputTokens,
+							cacheWriteTokens,
+							cacheReadTokens,
+						),
 					cancelReason,
 					streamingFailedMessage,
 				} satisfies ClineApiReqInfo)
@@ -3305,14 +3260,6 @@ export class Cline {
 					this.consecutiveMistakeCount++
 				}

-				telemetryService.capture({
-					event: "message sent",
-					properties: {
-						taskId: this.taskId,
-						chatMode: this.chatSettings.mode,
-					},
-				})
-
 				const recDidEndLoop = await this.recursivelyMakeClineRequests(this.userMessageContent)
 				didEndLoop = recDidEndLoop
 			} else {
@@ -3,7 +3,6 @@ import os from "os"
 import osName from "os-name"
 import { McpHub } from "../../services/mcp/McpHub"
 import { BrowserSettings } from "../../shared/BrowserSettings"
-import { DEFAULT_MCP_TIMEOUT_SECONDS } from "../../shared/mcp"

 export const SYSTEM_PROMPT = async (
 	cwd: string,
@@ -760,9 +759,8 @@ IMPORTANT: Regardless of what else you see in the MCP settings file, you must de
      "args": ["/path/to/weather-server/build/index.js"],
      "env": {
        "OPENWEATHER_API_KEY": "user-provided-api-key"
-      },
-      "timeout": ${DEFAULT_MCP_TIMEOUT_SECONDS}, // ${DEFAULT_MCP_TIMEOUT_SECONDS} seconds should be set as the default timeout for all MCP servers
-    }
+      }
+    },
  }
 }
 \`\`\`
@@ -8,7 +8,6 @@ import pWaitFor from "p-wait-for"
 import * as path from "path"
 import * as vscode from "vscode"
 import { buildApiHandler } from "../../api"
-import CheckpointTracker from "../../integrations/checkpoints/CheckpointTracker"
 import { downloadTask } from "../../integrations/misc/export-markdown"
 import { openFile, openImage } from "../../integrations/misc/open-file"
 import { fetchOpenGraphData, isImageUrl } from "../../integrations/misc/link-preview"
@@ -35,6 +34,7 @@ import { getNonce } from "./getNonce"
 import { getUri } from "./getUri"
 import { telemetryService } from "../../services/telemetry/TelemetryService"
 import { TelemetrySetting } from "../../shared/TelemetrySetting"
+import { validateThinkingBudget } from "../../utils/validation"

 /*
 https://github.com/microsoft/vscode-webview-ui-toolkit-samples/blob/main/default/weather-webview/src/providers/WeatherViewProvider.ts
@@ -59,12 +59,14 @@ type SecretKey =
 	| "liteLlmApiKey"
 	| "authToken"
 	| "authNonce"
+	| "asksageApiKey"
 	| "xaiApiKey"
 type GlobalStateKey =
 	| "apiProvider"
 	| "apiModelId"
 	| "awsRegion"
 	| "awsUseCrossRegionInference"
+	| "awsBedrockUsePromptCache"
 	| "awsProfile"
 	| "awsUseProfile"
 	| "vertexProjectId"
@@ -77,6 +79,7 @@ type GlobalStateKey =
 	| "openAiModelInfo"
 	| "ollamaModelId"
 	| "ollamaBaseUrl"
+	| "ollamaApiOptionsCtxNum"
 	| "lmStudioModelId"
 	| "lmStudioBaseUrl"
 	| "anthropicBaseUrl"
@@ -90,21 +93,22 @@ type GlobalStateKey =
 	| "userInfo"
 	| "previousModeApiProvider"
 	| "previousModeModelId"
+	| "previousModeThinkingBudgetTokens"
 	| "previousModeModelInfo"
 	| "liteLlmBaseUrl"
 	| "liteLlmModelId"
 	| "qwenApiLine"
 	| "requestyModelId"
-	| "requestyModelInfo"
 	| "togetherModelId"
 	| "mcpMarketplaceCatalog"
 	| "telemetrySetting"
+	| "asksageApiUrl"
+	| "thinkingBudgetTokens"

 export const GlobalFileNames = {
 	apiConversationHistory: "api_conversation_history.json",
 	uiMessages: "ui_messages.json",
 	openRouterModels: "openrouter_models.json",
-	requestyModels: "requesty_models.json",
 	mcpSettings: "cline_mcp_settings.json",
 	clineRules: ".clinerules",
 }
@@ -468,7 +472,7 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 					<meta http-equiv="Content-Security-Policy" content="${csp.join("; ")}">
 					<link rel="stylesheet" type="text/css" href="${stylesUri}">
 					<link href="${codiconsUri}" rel="stylesheet" />
-					<title>Roo Code</title>
+					<title>Cline</title>
 				</head>
 				<body>
 					<div id="root"></div>
@@ -499,7 +503,7 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 							}),
 						)
 						// post last cached models in case the call to endpoint fails
-						this.readDynamicProviderModels(GlobalFileNames.openRouterModels).then((openRouterModels) => {
+						this.readOpenRouterModels().then((openRouterModels) => {
 							if (openRouterModels) {
 								this.postMessageToWebview({
 									type: "openRouterModels",
@@ -541,34 +545,6 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 							const isOptedIn = telemetrySetting === "enabled"
 							telemetryService.updateTelemetryState(isOptedIn)
 						})
-
-						// post last cached models in case the call to endpoint fails
-						this.readDynamicProviderModels(GlobalFileNames.requestyModels).then((requestyModels) => {
-							if (requestyModels) {
-								this.postMessageToWebview({
-									type: "requestyModels",
-									requestyModels,
-								})
-							}
-						})
-
-						// gui relies on model info to be up-to-date to provide the most accurate pricing, so we need to fetch the latest details on launch.
-						// we do this for all users since many users switch between api providers and if they were to switch back to openrouter it would be showing outdated model info if we hadn't retrieved the latest at this point
-						// (see normalizeApiConfiguration > openrouter)
-						this.refreshRequestyModels().then(async (requestyModels) => {
-							if (requestyModels) {
-								// update model info in state (this needs to be done here since we don't want to update state while settings is open, and we may refresh models there)
-								const { apiConfiguration } = await this.getState()
-								if (apiConfiguration.requestyModelId) {
-									await this.updateGlobalState(
-										"requestyModelInfo",
-										requestyModels[apiConfiguration.requestyModelId],
-									)
-									await this.postStateToWebview()
-								}
-							}
-						})
-
 						break
 					case "newTask":
 						// Code that should run in response to the hello message command
@@ -593,6 +569,7 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 								awsSessionToken,
 								awsRegion,
 								awsUseCrossRegionInference,
+								awsBedrockUsePromptCache,
 								awsProfile,
 								awsUseProfile,
 								vertexProjectId,
@@ -603,6 +580,7 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 								openAiModelInfo,
 								ollamaModelId,
 								ollamaBaseUrl,
+								ollamaApiOptionsCtxNum,
 								lmStudioModelId,
 								lmStudioBaseUrl,
 								anthropicBaseUrl,
@@ -611,7 +589,6 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 								deepSeekApiKey,
 								requestyApiKey,
 								requestyModelId,
-								requestyModelInfo,
 								togetherApiKey,
 								togetherModelId,
 								qwenApiKey,
@@ -624,7 +601,10 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 								liteLlmModelId,
 								liteLlmApiKey,
 								qwenApiLine,
+								asksageApiKey,
+								asksageApiUrl,
 								xaiApiKey,
+								thinkingBudgetTokens,
 							} = message.apiConfiguration
 							await this.updateGlobalState("apiProvider", apiProvider)
 							await this.updateGlobalState("apiModelId", apiModelId)
@@ -635,6 +615,7 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 							await this.storeSecret("awsSessionToken", awsSessionToken)
 							await this.updateGlobalState("awsRegion", awsRegion)
 							await this.updateGlobalState("awsUseCrossRegionInference", awsUseCrossRegionInference)
+							await this.updateGlobalState("awsBedrockUsePromptCache", awsBedrockUsePromptCache)
 							await this.updateGlobalState("awsProfile", awsProfile)
 							await this.updateGlobalState("awsUseProfile", awsUseProfile)
 							await this.updateGlobalState("vertexProjectId", vertexProjectId)
@@ -645,6 +626,7 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 							await this.updateGlobalState("openAiModelInfo", openAiModelInfo)
 							await this.updateGlobalState("ollamaModelId", ollamaModelId)
 							await this.updateGlobalState("ollamaBaseUrl", ollamaBaseUrl)
+							await this.updateGlobalState("ollamaApiOptionsCtxNum", ollamaApiOptionsCtxNum)
 							await this.updateGlobalState("lmStudioModelId", lmStudioModelId)
 							await this.updateGlobalState("lmStudioBaseUrl", lmStudioBaseUrl)
 							await this.updateGlobalState("anthropicBaseUrl", anthropicBaseUrl)
@@ -665,8 +647,10 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 							await this.updateGlobalState("liteLlmModelId", liteLlmModelId)
 							await this.updateGlobalState("qwenApiLine", qwenApiLine)
 							await this.updateGlobalState("requestyModelId", requestyModelId)
-							await this.updateGlobalState("requestyModelInfo", requestyModelInfo)
 							await this.updateGlobalState("togetherModelId", togetherModelId)
+							await this.storeSecret("asksageApiKey", asksageApiKey)
+							await this.updateGlobalState("asksageApiUrl", asksageApiUrl)
+							await this.updateGlobalState("thinkingBudgetTokens", thinkingBudgetTokens)
 							if (this.cline) {
 								this.cline.api = buildApiHandler(message.apiConfiguration)
 							}
@@ -762,9 +746,6 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 					case "refreshOpenRouterModels":
 						await this.refreshOpenRouterModels()
 						break
-					case "refreshRequestyModels":
-						await this.refreshRequestyModels()
-						break
 					case "refreshOpenAiModels":
 						const { apiConfiguration } = await this.getState()
 						const openAiModels = await this.getOpenAiModels(
@@ -1025,25 +1006,24 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 			previousModeApiProvider: newApiProvider,
 			previousModeModelId: newModelId,
 			previousModeModelInfo: newModelInfo,
+			previousModeThinkingBudgetTokens: newThinkingBudgetTokens,
 		} = await this.getState()

 		// Save the last model used in this mode
 		await this.updateGlobalState("previousModeApiProvider", apiConfiguration.apiProvider)
+		await this.updateGlobalState("previousModeThinkingBudgetTokens", apiConfiguration.thinkingBudgetTokens)
 		switch (apiConfiguration.apiProvider) {
 			case "anthropic":
 			case "bedrock":
 			case "vertex":
 			case "gemini":
+			case "asksage":
 				await this.updateGlobalState("previousModeModelId", apiConfiguration.apiModelId)
 				break
 			case "openrouter":
 				await this.updateGlobalState("previousModeModelId", apiConfiguration.openRouterModelId)
 				await this.updateGlobalState("previousModeModelInfo", apiConfiguration.openRouterModelInfo)
 				break
-			case "requesty":
-				await this.updateGlobalState("previousModeModelId", apiConfiguration.requestyModelId)
-				await this.updateGlobalState("previousModeModelInfo", apiConfiguration.requestyModelInfo)
-				break
 			case "vscode-lm":
 				await this.updateGlobalState("previousModeModelId", apiConfiguration.vsCodeLmModelSelector)
 				break
@@ -1060,26 +1040,27 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 			case "litellm":
 				await this.updateGlobalState("previousModeModelId", apiConfiguration.liteLlmModelId)
 				break
+			case "requesty":
+				await this.updateGlobalState("previousModeModelId", apiConfiguration.requestyModelId)
+				break
 		}

 		// Restore the model used in previous mode
 		if (newApiProvider && newModelId) {
 			await this.updateGlobalState("apiProvider", newApiProvider)
+			await this.updateGlobalState("thinkingBudgetTokens", newThinkingBudgetTokens)
 			switch (newApiProvider) {
 				case "anthropic":
 				case "bedrock":
 				case "vertex":
 				case "gemini":
+				case "asksage":
 					await this.updateGlobalState("apiModelId", newModelId)
 					break
 				case "openrouter":
 					await this.updateGlobalState("openRouterModelId", newModelId)
 					await this.updateGlobalState("openRouterModelInfo", newModelInfo)
 					break
-				case "requesty":
-					await this.updateGlobalState("requestyModelId", newModelId)
-					await this.updateGlobalState("requestyModelInfo", newModelInfo)
-					break
 				case "vscode-lm":
 					await this.updateGlobalState("vsCodeLmModelSelector", newModelId)
 					break
@@ -1096,6 +1077,9 @@ export class ClineProvider implements vscode.WebviewViewProvider {
 				case "litellm":
 					await this.updateGlobalState("liteLlmModelId", newModelId)
 					break
+				case "requesty":
+					await this.updateGlobalState("requestyModelId", newModelId)
+					break
 			}

 			if (this.cline) {
@@ -1196,21 +1180,38 @@ export class ClineProvider implements vscode.WebviewViewProvider {

 	async getDocumentsPath(): Promise<string> {
 		if (process.platform === "win32") {
-			// If the user is running Win 7/Win Server 2008 r2+, we want to get the correct path to their Documents directory.
 			try {
 				const { stdout: docsPath } = await execa("powershell", [
 					"-NoProfile", // Ignore user's PowerShell profile(s)
 					"-Command",
 					"[System.Environment]::GetFolderPath([System.Environment+SpecialFolder]::MyDocuments)",
 				])
-				return docsPath.trim()
+				const trimmedPath = docsPath.trim()
+				if (trimmedPath) {
+					return trimmedPath
+				}
 			} catch (err) {
 				console.error("Failed to retrieve Windows Documents path. Falling back to homedir/Documents.")
-				return path.join(os.homedir(), "Documents")
 			}
-		} else {
-			return path.join(os.homedir(), "Documents") // On POSIX (macOS, Linux, etc.), assume ~/Documents by default (existing behavior, but may want to implement similar logic here)
+		} else if (process.platform === "linux") {
+			try {
+				// First check if xdg-user-dir exists
+				await execa("which", ["xdg-user-dir"])
+
+				// If it exists, try to get XDG documents path
+				const { stdout } = await execa("xdg-user-dir", ["DOCUMENTS"])
+				const trimmedPath = stdout.trim()
+				if (trimmedPath) {
+					return trimmedPath
+				}
+			} catch {
+				// Log error but continue to fallback
+				console.error("Failed to retrieve XDG Documents path. Falling back to homedir/Documents.")
+			}
 		}
+
+		// Default fallback for all platforms
+		return path.join(os.homedir(), "Documents")
 	}

 	async ensureMcpServersDirectoryExists(): Promise<string> {
@@ -1535,61 +1536,16 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 		return cacheDir
 	}

-	async readDynamicProviderModels(filename: string): Promise<Record<string, ModelInfo> | undefined> {
-		const filePath = path.join(await this.ensureCacheDirectoryExists(), filename)
-		const fileExists = await fileExistsAtPath(filePath)
+	async readOpenRouterModels(): Promise<Record<string, ModelInfo> | undefined> {
+		const openRouterModelsFilePath = path.join(await this.ensureCacheDirectoryExists(), GlobalFileNames.openRouterModels)
+		const fileExists = await fileExistsAtPath(openRouterModelsFilePath)
 		if (fileExists) {
-			const fileContents = await fs.readFile(filePath, "utf8")
+			const fileContents = await fs.readFile(openRouterModelsFilePath, "utf8")
 			return JSON.parse(fileContents)
 		}
 		return undefined
 	}

-	adjustPriceToMillionTokens(price: any) {
-		if (price) {
-			return parseFloat(price) * 1_000_000
-		}
-		return undefined
-	}
-
-	async refreshRequestyModels() {
-		const requestyModelsFilePath = path.join(await this.ensureCacheDirectoryExists(), GlobalFileNames.requestyModels)
-
-		let models: Record<string, ModelInfo> = {}
-		try {
-			const response = await axios.get("https://router.requesty.ai/v1/models")
-			if (response.data?.data) {
-				for (const model of response.data.data) {
-					const modelInfo: ModelInfo = {
-						maxTokens: model.max_output_tokens,
-						contextWindow: model.context_window,
-						supportsImages: model.supports_images || undefined,
-						supportsComputerUse: model.supports_computer_use || undefined,
-						supportsPromptCache: model.supports_caching || undefined,
-						inputPrice: this.adjustPriceToMillionTokens(model.input_price),
-						outputPrice: this.adjustPriceToMillionTokens(model.output_price),
-						cacheWritesPrice: this.adjustPriceToMillionTokens(model.caching_price),
-						cacheReadsPrice: this.adjustPriceToMillionTokens(model.cached_price),
-						description: model.description,
-					}
-					models[model.id] = modelInfo
-				}
-				await fs.writeFile(requestyModelsFilePath, JSON.stringify(models))
-				console.log("Requesty models fetched and saved", models)
-			} else {
-				console.error("Invalid response from Requesty API")
-			}
-		} catch (error) {
-			console.error("Error fetching Requesty models:", error)
-		}
-
-		await this.postMessageToWebview({
-			type: "requestyModels",
-			requestyModels: models,
-		})
-		return models
-	}
-
 	async refreshOpenRouterModels() {
 		const openRouterModelsFilePath = path.join(await this.ensureCacheDirectoryExists(), GlobalFileNames.openRouterModels)

@@ -1624,14 +1580,20 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			*/
 			if (response.data?.data) {
 				const rawModels = response.data.data
+				const parsePrice = (price: any) => {
+					if (price) {
+						return parseFloat(price) * 1_000_000
+					}
+					return undefined
+				}
 				for (const rawModel of rawModels) {
 					const modelInfo: ModelInfo = {
 						maxTokens: rawModel.top_provider?.max_completion_tokens,
 						contextWindow: rawModel.context_length,
 						supportsImages: rawModel.architecture?.modality?.includes("image"),
 						supportsPromptCache: false,
-						inputPrice: this.adjustPriceToMillionTokens(rawModel.pricing?.prompt),
-						outputPrice: this.adjustPriceToMillionTokens(rawModel.pricing?.completion),
+						inputPrice: parsePrice(rawModel.pricing?.prompt),
+						outputPrice: parsePrice(rawModel.pricing?.completion),
 						description: rawModel.description,
 					}

@@ -1640,6 +1602,7 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 						case "anthropic/claude-3-7-sonnet:beta":
 						case "anthropic/claude-3.7-sonnet":
 						case "anthropic/claude-3.7-sonnet:beta":
+						case "anthropic/claude-3.7-sonnet:thinking":
 						case "anthropic/claude-3.5-sonnet":
 						case "anthropic/claude-3.5-sonnet:beta":
 							// NOTE: this needs to be synced with api.ts/openrouter default model info
@@ -1765,20 +1728,6 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont

 		const { taskDirPath, apiConversationHistoryFilePath, uiMessagesFilePath } = await this.getTaskWithId(id)

-		// Delete checkpoints
-		// deleteCheckpoints will determine if the task has legacy checkpoints or not and handle it accordingly
-		console.info("deleting checkpoints")
-		const taskHistory = ((await this.getGlobalState("taskHistory")) as HistoryItem[] | undefined) || []
-		const historyItem = taskHistory.find((item) => item.id === id)
-		//console.log("historyItem: ", historyItem)
-		if (historyItem) {
-			try {
-				await CheckpointTracker.deleteCheckpoints(id, historyItem, this.context.globalStorageUri.fsPath)
-			} catch (error) {
-				console.error(`Failed to delete checkpoints for task ${id}:`, error)
-			}
-		}
-
 		await this.deleteTaskFromState(id)

 		// Delete the task files
@@ -1795,12 +1744,21 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			await fs.unlink(legacyMessagesFilePath)
 		}

+		// Delete the checkpoints directory if it exists
+		const checkpointsDir = path.join(taskDirPath, "checkpoints")
+		if (await fileExistsAtPath(checkpointsDir)) {
+			try {
+				await fs.rm(checkpointsDir, { recursive: true, force: true })
+			} catch (error) {
+				console.error(`Failed to delete checkpoints directory for task ${id}:`, error)
+				// Continue with deletion of task directory - don't throw since this is a cleanup operation
+			}
+		}
+
 		await fs.rmdir(taskDirPath) // succeeds if the dir is empty
 	}

 	async deleteTaskFromState(id: string) {
-		console.log("deleteTaskFromState: ", id)
-
 		// Remove the task from history
 		const taskHistory = ((await this.getGlobalState("taskHistory")) as HistoryItem[] | undefined) || []
 		const updatedTaskHistory = taskHistory.filter((task) => task.id !== id)
@@ -1871,7 +1829,7 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont

 	/*
 	It seems that some API messages do not comply with vscode state requirements. Either the Anthropic library is manipulating these values somehow in the backend in a way thats creating cyclic references, or the API returns a function or a Symbol as part of the message content.
-	VSCode docs about state: "The value must be JSON-stringifyable ... value  A value. MUST not contain cyclic references."
+	VSCode docs about state: "The value must be JSON-stringifyable ... value — A value. MUST not contain cyclic references."
 	For now we'll store the conversation history in memory, and if we need to store in state directly we'd need to do a manual conversion to ensure proper json stringification.
 	*/

@@ -1914,6 +1872,7 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			awsSessionToken,
 			awsRegion,
 			awsUseCrossRegionInference,
+			awsBedrockUsePromptCache,
 			awsProfile,
 			awsUseProfile,
 			vertexProjectId,
@@ -1924,6 +1883,7 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			openAiModelInfo,
 			ollamaModelId,
 			ollamaBaseUrl,
+			ollamaApiOptionsCtxNum,
 			lmStudioModelId,
 			lmStudioBaseUrl,
 			anthropicBaseUrl,
@@ -1932,7 +1892,6 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			deepSeekApiKey,
 			requestyApiKey,
 			requestyModelId,
-			requestyModelInfo,
 			togetherApiKey,
 			togetherModelId,
 			qwenApiKey,
@@ -1954,10 +1913,14 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			previousModeApiProvider,
 			previousModeModelId,
 			previousModeModelInfo,
+			previousModeThinkingBudgetTokens,
 			qwenApiLine,
 			liteLlmApiKey,
 			telemetrySetting,
+			asksageApiKey,
+			asksageApiUrl,
 			xaiApiKey,
+			thinkingBudgetTokens,
 		] = await Promise.all([
 			this.getGlobalState("apiProvider") as Promise<ApiProvider | undefined>,
 			this.getGlobalState("apiModelId") as Promise<string | undefined>,
@@ -1968,6 +1931,7 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			this.getSecret("awsSessionToken") as Promise<string | undefined>,
 			this.getGlobalState("awsRegion") as Promise<string | undefined>,
 			this.getGlobalState("awsUseCrossRegionInference") as Promise<boolean | undefined>,
+			this.getGlobalState("awsBedrockUsePromptCache") as Promise<boolean | undefined>,
 			this.getGlobalState("awsProfile") as Promise<string | undefined>,
 			this.getGlobalState("awsUseProfile") as Promise<boolean | undefined>,
 			this.getGlobalState("vertexProjectId") as Promise<string | undefined>,
@@ -1978,6 +1942,7 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			this.getGlobalState("openAiModelInfo") as Promise<ModelInfo | undefined>,
 			this.getGlobalState("ollamaModelId") as Promise<string | undefined>,
 			this.getGlobalState("ollamaBaseUrl") as Promise<string | undefined>,
+			this.getGlobalState("ollamaApiOptionsCtxNum") as Promise<string | undefined>,
 			this.getGlobalState("lmStudioModelId") as Promise<string | undefined>,
 			this.getGlobalState("lmStudioBaseUrl") as Promise<string | undefined>,
 			this.getGlobalState("anthropicBaseUrl") as Promise<string | undefined>,
@@ -1986,7 +1951,6 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			this.getSecret("deepSeekApiKey") as Promise<string | undefined>,
 			this.getSecret("requestyApiKey") as Promise<string | undefined>,
 			this.getGlobalState("requestyModelId") as Promise<string | undefined>,
-			this.getGlobalState("requestyModelInfo") as Promise<ModelInfo | undefined>,
 			this.getSecret("togetherApiKey") as Promise<string | undefined>,
 			this.getGlobalState("togetherModelId") as Promise<string | undefined>,
 			this.getSecret("qwenApiKey") as Promise<string | undefined>,
@@ -2008,10 +1972,14 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			this.getGlobalState("previousModeApiProvider") as Promise<ApiProvider | undefined>,
 			this.getGlobalState("previousModeModelId") as Promise<string | undefined>,
 			this.getGlobalState("previousModeModelInfo") as Promise<ModelInfo | undefined>,
+			this.getGlobalState("previousModeThinkingBudgetTokens") as Promise<number | undefined>,
 			this.getGlobalState("qwenApiLine") as Promise<string | undefined>,
 			this.getSecret("liteLlmApiKey") as Promise<string | undefined>,
 			this.getGlobalState("telemetrySetting") as Promise<TelemetrySetting | undefined>,
+			this.getSecret("asksageApiKey") as Promise<string | undefined>,
+			this.getGlobalState("asksageApiUrl") as Promise<string | undefined>,
 			this.getSecret("xaiApiKey") as Promise<string | undefined>,
+			this.getGlobalState("thinkingBudgetTokens") as Promise<number | undefined>,
 		])

 		let apiProvider: ApiProvider
@@ -2045,6 +2013,7 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 				awsSessionToken,
 				awsRegion,
 				awsUseCrossRegionInference,
+				awsBedrockUsePromptCache,
 				awsProfile,
 				awsUseProfile,
 				vertexProjectId,
@@ -2055,6 +2024,7 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 				openAiModelInfo,
 				ollamaModelId,
 				ollamaBaseUrl,
+				ollamaApiOptionsCtxNum,
 				lmStudioModelId,
 				lmStudioBaseUrl,
 				anthropicBaseUrl,
@@ -2063,7 +2033,6 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 				deepSeekApiKey,
 				requestyApiKey,
 				requestyModelId,
-				requestyModelInfo,
 				togetherApiKey,
 				togetherModelId,
 				qwenApiKey,
@@ -2074,9 +2043,12 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 				openRouterModelInfo,
 				vsCodeLmModelSelector,
 				o3MiniReasoningEffort,
+				thinkingBudgetTokens,
 				liteLlmBaseUrl,
 				liteLlmModelId,
 				liteLlmApiKey,
+				asksageApiKey,
+				asksageApiUrl,
 				xaiApiKey,
 			},
 			lastShownAnnouncementId,
@@ -2090,6 +2062,7 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			previousModeApiProvider,
 			previousModeModelId,
 			previousModeModelInfo,
+			previousModeThinkingBudgetTokens,
 			mcpMarketplaceEnabled,
 			telemetrySetting: telemetrySetting || "unset",
 		}
@@ -2221,6 +2194,7 @@ Here is the project's README to help you get started:\n\n${mcpDetails.readmeCont
 			"mistralApiKey",
 			"liteLlmApiKey",
 			"authToken",
+			"asksageApiKey",
 			"xaiApiKey",
 		]
 		for (const key of secretKeys) {
@@ -0,0 +1,453 @@
+import fs from "fs/promises"
+import * as path from "path"
+import simpleGit, { SimpleGit } from "simple-git"
+import * as vscode from "vscode"
+import { HistoryItem } from "../../shared/HistoryItem"
+import { GitOperations } from "./CheckpointGitOperations"
+import { getShadowGitPath, hashWorkingDir, getWorkingDirectory, detectLegacyCheckpoint } from "./CheckpointUtils"
+
+/**
+ * CheckpointTracker Module
+ *
+ * Core implementation of Cline's Checkpoints system that provides version control
+ * capabilities without interfering with the user's main Git repository. Key features:
+ *
+ * Shadow Git Repository:
+ * - Creates and manages an isolated Git repository for tracking checkpoints
+ * - Handles nested Git repositories by temporarily disabling them
+ * - Configures Git settings automatically (identity, LFS, etc.)
+ *
+ * File Management:
+ * - Integrates with CheckpointExclusions for file filtering
+ * - Handles workspace validation and path resolution
+ * - Manages Git worktree configuration
+ *
+ * Checkpoint Operations:
+ * - Creates checkpoints (commits) of the current state
+ * - Provides diff capabilities between checkpoints
+ * - Supports resetting to previous checkpoints
+ *
+ * Safety Features:
+ * - Prevents usage in sensitive directories (home, desktop, etc.)
+ * - Validates workspace configuration
+ * - Handles cleanup and resource disposal
+ *
+ * Checkpoint Architecture:
+ * - Uses a branch-per-task model to consolidate shadow git repositories
+ * - Each task gets its own branch within a single shadow git per workspace
+ * - Maintains backward compatibility with legacy checkpoint structure
+ * - Automatically cleans up by deleting task branches when tasks are removed
+ */
+
+class CheckpointTracker {
+	private globalStoragePath: string
+	private taskId: string
+	private cwd: string
+	private cwdHash: string
+	private lastRetrievedShadowGitConfigWorkTree?: string
+	private lastCheckpointHash?: string
+	private isLegacyCheckpoint: boolean = false
+	private gitOperations: GitOperations
+
+	/**
+	 * Creates a new CheckpointTracker instance to manage checkpoints for a specific task.
+	 * The constructor is private - use the static create() method to instantiate.
+	 *
+	 * @param taskId - Unique identifier for the task being tracked
+	 * @param cwd - The current working directory to track files in
+	 * @param cwdHash - Hash of the working directory path for shadow git organization
+	 */
+	private constructor(globalStoragePath: string, taskId: string, cwd: string, cwdHash: string) {
+		this.globalStoragePath = globalStoragePath
+		this.taskId = taskId
+		this.cwd = cwd
+		this.cwdHash = cwdHash
+		this.gitOperations = new GitOperations(cwd, false) // Initialize with non-legacy mode
+	}
+
+	/**
+	 * Creates a new CheckpointTracker instance for tracking changes in a task.
+	 * Handles initialization of the shadow git repository and branch setup.
+	 *
+	 * @param taskId - Unique identifier for the task to track
+	 * @param globalStoragePath - the globalStorage path
+	 * @returns Promise resolving to new CheckpointTracker instance, or undefined if checkpoints are disabled
+	 * @throws Error if:
+	 * - globalStoragePath is not supplied
+	 * - Git is not installed
+	 * - Working directory is invalid or in a protected location
+	 * - Shadow git initialization fails
+	 *
+	 * Key operations:
+	 * - Validates git installation and settings
+	 * - Creates/initializes shadow git repository
+	 * - Detects and handles legacy checkpoint structure
+	 * - Sets up task-specific branch for new checkpoints
+	 *
+	 * Configuration:
+	 * - Respects 'cline.enableCheckpoints' VS Code setting
+	 * - Uses branch-per-task architecture for new checkpoints
+	 * - Maintains backwards compatibility with legacy structure
+	 */
+	public static async create(taskId: string, globalStoragePath: string | undefined): Promise<CheckpointTracker | undefined> {
+		if (!globalStoragePath) {
+			throw new Error("Global storage path is required to create a checkpoint tracker")
+		}
+		try {
+			console.info(`Creating new CheckpointTracker for task ${taskId}`)
+
+			// Check if checkpoints are disabled in VS Code settings
+			const enableCheckpoints = vscode.workspace.getConfiguration("cline").get<boolean>("enableCheckpoints") ?? true
+			if (!enableCheckpoints) {
+				return undefined // Don't create tracker when disabled
+			}
+
+			// Check if git is installed by attempting to get version
+			try {
+				await simpleGit().version()
+			} catch (error) {
+				throw new Error("Git must be installed to use checkpoints.") // FIXME: must match what we check for in TaskHeader to show link
+			}
+
+			const workingDir = await getWorkingDirectory()
+			const cwdHash = hashWorkingDir(workingDir)
+			console.debug(`Repository ID (cwdHash): ${cwdHash}`)
+
+			const newTracker = new CheckpointTracker(globalStoragePath, taskId, workingDir, cwdHash)
+
+			// Check if this is a legacy task
+			newTracker.isLegacyCheckpoint = await detectLegacyCheckpoint(newTracker.globalStoragePath, newTracker.taskId)
+			if (newTracker.isLegacyCheckpoint) {
+				console.debug("Using legacy checkpoint path structure")
+				const gitPath = await getShadowGitPath(
+					newTracker.globalStoragePath,
+					newTracker.taskId,
+					newTracker.cwdHash,
+					newTracker.isLegacyCheckpoint,
+				)
+				await GitOperations.initShadowGit(gitPath, workingDir, newTracker.isLegacyCheckpoint)
+				await newTracker.gitOperations.switchToTaskBranch(newTracker.taskId, gitPath)
+				return newTracker
+			}
+
+			// Branch-per-task structure
+			const gitPath = await getShadowGitPath(
+				newTracker.globalStoragePath,
+				newTracker.taskId,
+				newTracker.cwdHash,
+				newTracker.isLegacyCheckpoint,
+			)
+			await GitOperations.initShadowGit(gitPath, workingDir, newTracker.isLegacyCheckpoint)
+			await newTracker.gitOperations.switchToTaskBranch(newTracker.taskId, gitPath)
+			return newTracker
+		} catch (error) {
+			console.error("Failed to create CheckpointTracker:", error)
+			throw error
+		}
+	}
+
+	/**
+	 * Creates a new checkpoint commit in the shadow git repository.
+	 *
+	 * Key behaviors:
+	 * - Creates commit with checkpoint files in shadow git repo
+	 * - Handles both legacy and branch-per-task checkpoint structures
+	 * - For new tasks, switches to task-specific branch first
+	 * - Caches the created commit hash
+	 *
+	 * Commit structure:
+	 * - Legacy: Simple "checkpoint" message
+	 * - Branch-per-task: "checkpoint-{cwdHash}-{taskId}"
+	 * - Always allows empty commits
+	 *
+	 * Dependencies:
+	 * - Requires initialized shadow git (getShadowGitPath)
+	 * - For new checkpoints, requires task branch setup
+	 * - Uses addCheckpointFiles to stage changes
+	 *
+	 * @returns Promise<string | undefined> The created commit hash, or undefined if:
+	 * - Shadow git access fails
+	 * - Branch switch fails
+	 * - Staging files fails
+	 * - Commit creation fails
+	 * @throws Error if unable to:
+	 * - Access shadow git path
+	 * - Initialize simple-git
+	 * - Switch branches
+	 * - Stage or commit files
+	 */
+	public async commit(): Promise<string | undefined> {
+		try {
+			console.info(`Creating new checkpoint commit for task ${this.taskId}`)
+			const gitPath = await getShadowGitPath(this.globalStoragePath, this.taskId, this.cwdHash, this.isLegacyCheckpoint)
+			const git = simpleGit(path.dirname(gitPath))
+
+			console.info(`Using shadow git at: ${gitPath}`)
+
+			// Disable nested git repos before any operations
+			await this.gitOperations.renameNestedGitRepos(true)
+
+			try {
+				if (!this.isLegacyCheckpoint) {
+					await this.gitOperations.switchToTaskBranch(this.taskId, gitPath)
+				}
+				await this.gitOperations.addCheckpointFiles(git, gitPath)
+
+				const commitMessage = this.isLegacyCheckpoint ? "checkpoint" : "checkpoint-" + this.cwdHash + "-" + this.taskId
+
+				console.info(
+					`Creating ${this.isLegacyCheckpoint ? "legacy" : "new"} checkpoint commit with message: ${commitMessage}`,
+				)
+				const result = await git.commit(commitMessage, {
+					"--allow-empty": null,
+				})
+				const commitHash = result.commit || ""
+				this.lastCheckpointHash = commitHash
+				console.warn(`Checkpoint commit created.`)
+				return commitHash
+			} finally {
+				// Always re-enable nested git repos
+				await this.gitOperations.renameNestedGitRepos(false)
+			}
+		} catch (error) {
+			console.error("Failed to create checkpoint:", {
+				taskId: this.taskId,
+				error,
+				isLegacyCheckpoint: this.isLegacyCheckpoint,
+			})
+			throw new Error(`Failed to create checkpoint: ${error instanceof Error ? error.message : String(error)}`)
+		}
+	}
+
+	/**
+	 * Retrieves the worktree path from the shadow git configuration.
+	 * The worktree path indicates where the shadow git repository is tracking files,
+	 * which should match the current workspace directory.
+	 *
+	 * Key behaviors:
+	 * - Caches result in lastRetrievedShadowGitConfigWorkTree to avoid repeated reads
+	 * - Returns cached value if available
+	 * - Reads git config if no cached value exists
+	 * - Handles both legacy and new checkpoint structures
+	 *
+	 * Configuration read:
+	 * - Uses simple-git to read core.worktree config
+	 * - Operates on shadow git at path from getShadowGitPath()
+	 *
+	 * @returns Promise<string | undefined> The configured worktree path, or undefined if:
+	 * - Shadow git repository doesn't exist
+	 * - Config read fails
+	 * - No worktree is configured
+	 * @throws Error if unable to:
+	 * - Access shadow git path
+	 * - Initialize simple-git
+	 * - Read git configuration
+	 */
+	public async getShadowGitConfigWorkTree(): Promise<string | undefined> {
+		if (this.lastRetrievedShadowGitConfigWorkTree) {
+			return this.lastRetrievedShadowGitConfigWorkTree
+		}
+		try {
+			const gitPath = await getShadowGitPath(this.globalStoragePath, this.taskId, this.cwdHash, this.isLegacyCheckpoint)
+			this.lastRetrievedShadowGitConfigWorkTree = await this.gitOperations.getShadowGitConfigWorkTree(gitPath)
+			return this.lastRetrievedShadowGitConfigWorkTree
+		} catch (error) {
+			console.error("Failed to get shadow git config worktree:", error)
+			return undefined
+		}
+	}
+
+	/**
+	 * Resets the shadow git repository's HEAD to a specific checkpoint commit.
+	 * This will discard all changes after the target commit and restore the
+	 * working directory to that checkpoint's state.
+	 *
+	 * Dependencies:
+	 * - Requires initialized shadow git (getShadowGitPath)
+	 * - For new checkpoints, requires task branch setup
+	 * - Must be called with a valid commit hash from this task's history
+	 *
+	 * @param commitHash - The hash of the checkpoint commit to reset to
+	 * @returns Promise<void> Resolves when reset is complete
+	 * @throws Error if unable to:
+	 * - Access shadow git path
+	 * - Initialize simple-git
+	 * - Switch to task branch
+	 * - Reset to target commit
+	 */
+	public async resetHead(commitHash: string): Promise<void> {
+		console.info(`Resetting to checkpoint: ${commitHash}`)
+		const gitPath = await getShadowGitPath(this.globalStoragePath, this.taskId, this.cwdHash, this.isLegacyCheckpoint)
+		const git = simpleGit(path.dirname(gitPath))
+		console.debug(`Using shadow git at: ${gitPath}`)
+		await this.gitOperations.switchToTaskBranch(this.taskId, gitPath)
+		await git.reset(["--hard", commitHash]) // Hard reset to target commit
+		console.debug(`Successfully reset to checkpoint: ${commitHash}`)
+	}
+
+	/**
+	 * Return an array describing changed files between one commit and either:
+	 *   - another commit, or
+	 *   - the current working directory (including uncommitted changes).
+	 *
+	 * If `rhsHash` is omitted, compares `lhsHash` to the working directory.
+	 * If you want truly untracked files to appear, `git add` them first.
+	 *
+	 * @param lhsHash - The commit to compare from (older commit)
+	 * @param rhsHash - The commit to compare to (newer commit).
+	 *                  If omitted, we compare to the working directory.
+	 * @returns Array of file changes with before/after content
+	 */
+	public async getDiffSet(
+		lhsHash?: string,
+		rhsHash?: string,
+	): Promise<
+		Array<{
+			relativePath: string
+			absolutePath: string
+			before: string
+			after: string
+		}>
+	> {
+		const gitPath = await getShadowGitPath(this.globalStoragePath, this.taskId, this.cwdHash, this.isLegacyCheckpoint)
+		const git = simpleGit(path.dirname(gitPath))
+
+		if (!this.isLegacyCheckpoint) {
+			await this.gitOperations.switchToTaskBranch(this.taskId, gitPath)
+		}
+
+		console.info(`Getting diff between commits: ${lhsHash || "initial"} -> ${rhsHash || "working directory"}`)
+
+		// If lhsHash is missing, use the initial commit of the repo
+		let baseHash = lhsHash
+		if (!baseHash) {
+			const rootCommit = await git.raw(["rev-list", "--max-parents=0", "HEAD"])
+			baseHash = rootCommit.trim()
+			console.debug(`Using root commit as base: ${baseHash}`)
+		}
+
+		// Stage all changes so that untracked files appear in diff summary
+		await this.gitOperations.addCheckpointFiles(git, gitPath)
+
+		const diffSummary = rhsHash ? await git.diffSummary([`${baseHash}..${rhsHash}`]) : await git.diffSummary([baseHash])
+		console.info(`Found ${diffSummary.files.length} changed files`)
+
+		// For each changed file, gather before/after content
+		const result = []
+		const cwdPath = (await this.getShadowGitConfigWorkTree()) || this.cwd || ""
+		const files = diffSummary.files.map((f) => f.file)
+		const batchSize = 50
+
+		// Get list of files that exist in base commit
+		const existingFiles = await this.getExistingFiles(git, baseHash, files)
+
+		// Process files in batches
+		for (let i = 0; i < files.length; i += batchSize) {
+			const batch = files.slice(i, i + batchSize)
+
+			// Split batch into existing and new files
+			const existingBatch = batch.filter((file) => existingFiles.has(file))
+			const newBatch = batch.filter((file) => !existingFiles.has(file))
+
+			// Get before contents for existing files
+			let beforeContents: string[] = new Array(batch.length).fill("")
+			if (existingBatch.length > 0) {
+				await git.addConfig("core.quotePath", "false")
+				await git.addConfig("core.precomposeunicode", "true")
+				const args = ["show", "--format="]
+				existingBatch.forEach((file) => {
+					args.push(`${baseHash}:${file}`)
+				})
+				const beforeResult = await git.raw(args)
+				const existingContents = beforeResult.split("\n\0\n")
+				// Map contents back to original batch positions
+				existingBatch.forEach((file, index) => {
+					const batchIndex = batch.indexOf(file)
+					if (batchIndex !== -1) {
+						beforeContents[batchIndex] = existingContents[index] || ""
+					}
+				})
+			}
+
+			// Get after contents
+			let afterContents: string[] = []
+			if (rhsHash) {
+				// Split after files into existing and new in target commit
+				const afterExistingFiles = await this.getExistingFiles(git, rhsHash, batch)
+				const afterExistingBatch = batch.filter((file) => afterExistingFiles.has(file))
+
+				if (afterExistingBatch.length > 0) {
+					const args = ["show", "--format="]
+					afterExistingBatch.forEach((file) => {
+						args.push(`${rhsHash}:${file}`)
+					})
+					const afterResult = await git.raw(args)
+					const existingContents = afterResult.split("\n\0\n")
+					afterContents = new Array(batch.length).fill("")
+					afterExistingBatch.forEach((file, index) => {
+						const batchIndex = batch.indexOf(file)
+						if (batchIndex !== -1) {
+							afterContents[batchIndex] = existingContents[index] || ""
+						}
+					})
+				}
+			} else {
+				// Read from disk for working directory changes
+				afterContents = await Promise.all(
+					batch.map(async (filePath) => {
+						try {
+							return await fs.readFile(path.join(cwdPath, filePath), "utf8")
+						} catch (_) {
+							return ""
+						}
+					}),
+				)
+			}
+
+			// Add results for this batch
+			for (let j = 0; j < batch.length; j++) {
+				const filePath = batch[j]
+				const absolutePath = path.join(cwdPath, filePath)
+				result.push({
+					relativePath: filePath,
+					absolutePath,
+					before: beforeContents[j] || "",
+					after: afterContents[j] || "",
+				})
+			}
+		}
+		return result
+	}
+
+	/**
+	 * Deletes all checkpoint data for a given task.
+	 * Handles both legacy checkpoints and branch-per-task checkpoints.
+	 *
+	 * @param taskId - The ID of the task whose checkpoints should be deleted
+	 * @param historyItem - The history item containing the shadow git config for this task
+	 * @param globalStoragePath - the globalStorage path
+	 * @throws Error if deletion fails
+	 */
+	public static async deleteCheckpoints(taskId: string, historyItem: HistoryItem, globalStoragePath: string): Promise<void> {
+		if (!globalStoragePath) {
+			throw new Error("Global storage uri is invalid")
+		}
+		await GitOperations.deleteTaskBranchStatic(taskId, historyItem, globalStoragePath)
+	}
+
+	/**
+	 * Helper function to get a set of files that exist in a given commit
+	 */
+	private async getExistingFiles(git: SimpleGit, commitHash: string, files: string[]): Promise<Set<string>> {
+		try {
+			const result = await git.raw(["ls-tree", "-r", "--name-only", commitHash])
+			const existingFiles = new Set<string>(result.split("\n"))
+			return existingFiles
+		} catch (error) {
+			console.error("Error getting existing files:", error)
+			return new Set()
+		}
+	}
+}
+
+export default CheckpointTracker
@@ -1,100 +1,31 @@
 import fs from "fs/promises"
+import os from "os"
 import * as path from "path"
 import simpleGit, { SimpleGit } from "simple-git"
 import * as vscode from "vscode"
-import { HistoryItem } from "../../shared/HistoryItem"
-import { GitOperations } from "./CheckpointGitOperations"
-import { getShadowGitPath, hashWorkingDir, getWorkingDirectory, detectLegacyCheckpoint } from "./CheckpointUtils"
-
-/**
- * CheckpointTracker Module
- *
- * Core implementation of Cline's Checkpoints system that provides version control
- * capabilities without interfering with the user's main Git repository. Key features:
- *
- * Shadow Git Repository:
- * - Creates and manages an isolated Git repository for tracking checkpoints
- * - Handles nested Git repositories by temporarily disabling them
- * - Configures Git settings automatically (identity, LFS, etc.)
- *
- * File Management:
- * - Integrates with CheckpointExclusions for file filtering
- * - Handles workspace validation and path resolution
- * - Manages Git worktree configuration
- *
- * Checkpoint Operations:
- * - Creates checkpoints (commits) of the current state
- * - Provides diff capabilities between checkpoints
- * - Supports resetting to previous checkpoints
- *
- * Safety Features:
- * - Prevents usage in sensitive directories (home, desktop, etc.)
- * - Validates workspace configuration
- * - Handles cleanup and resource disposal
- *
- * Checkpoint Architecture:
- * - Uses a branch-per-task model to consolidate shadow git repositories
- * - Each task gets its own branch within a single shadow git per workspace
- * - Maintains backward compatibility with legacy checkpoint structure
- * - Automatically cleans up by deleting task branches when tasks are removed
- */
+import { ClineProvider } from "../../core/webview/ClineProvider"
+import { fileExistsAtPath } from "../../utils/fs"
+import { globby } from "globby"

 class CheckpointTracker {
-	private globalStoragePath: string
+	private providerRef: WeakRef<ClineProvider>
 	private taskId: string
+	private disposables: vscode.Disposable[] = []
 	private cwd: string
-	private cwdHash: string
 	private lastRetrievedShadowGitConfigWorkTree?: string
-	private lastCheckpointHash?: string
-	private isLegacyCheckpoint: boolean = false
-	private gitOperations: GitOperations
+	lastCheckpointHash?: string

-	/**
-	 * Creates a new CheckpointTracker instance to manage checkpoints for a specific task.
-	 * The constructor is private - use the static create() method to instantiate.
-	 *
-	 * @param taskId - Unique identifier for the task being tracked
-	 * @param cwd - The current working directory to track files in
-	 * @param cwdHash - Hash of the working directory path for shadow git organization
-	 */
-	private constructor(globalStoragePath: string, taskId: string, cwd: string, cwdHash: string) {
-		this.globalStoragePath = globalStoragePath
+	private constructor(provider: ClineProvider, taskId: string, cwd: string) {
+		this.providerRef = new WeakRef(provider)
 		this.taskId = taskId
 		this.cwd = cwd
-		this.cwdHash = cwdHash
-		this.gitOperations = new GitOperations(cwd, false) // Initialize with non-legacy mode
 	}

-	/**
-	 * Creates a new CheckpointTracker instance for tracking changes in a task.
-	 * Handles initialization of the shadow git repository and branch setup.
-	 *
-	 * @param taskId - Unique identifier for the task to track
-	 * @param globalStoragePath - the globalStorage path
-	 * @returns Promise resolving to new CheckpointTracker instance, or undefined if checkpoints are disabled
-	 * @throws Error if:
-	 * - globalStoragePath is not supplied
-	 * - Git is not installed
-	 * - Working directory is invalid or in a protected location
-	 * - Shadow git initialization fails
-	 *
-	 * Key operations:
-	 * - Validates git installation and settings
-	 * - Creates/initializes shadow git repository
-	 * - Detects and handles legacy checkpoint structure
-	 * - Sets up task-specific branch for new checkpoints
-	 *
-	 * Configuration:
-	 * - Respects 'cline.enableCheckpoints' VS Code setting
-	 * - Uses branch-per-task architecture for new checkpoints
-	 * - Maintains backwards compatibility with legacy structure
-	 */
-	public static async create(taskId: string, globalStoragePath: string | undefined): Promise<CheckpointTracker | undefined> {
-		if (!globalStoragePath) {
-			throw new Error("Global storage path is required to create a checkpoint tracker")
-		}
+	public static async create(taskId: string, provider?: ClineProvider): Promise<CheckpointTracker | undefined> {
 		try {
-			console.info(`Creating new CheckpointTracker for task ${taskId}`)
+			if (!provider) {
+				throw new Error("Provider is required to create a checkpoint tracker")
+			}

 			// Check if checkpoints are disabled in VS Code settings
 			const enableCheckpoints = vscode.workspace.getConfiguration("cline").get<boolean>("enableCheckpoints") ?? true
@@ -109,36 +40,9 @@ class CheckpointTracker {
 				throw new Error("Git must be installed to use checkpoints.") // FIXME: must match what we check for in TaskHeader to show link
 			}

-			const workingDir = await getWorkingDirectory()
-			const cwdHash = hashWorkingDir(workingDir)
-			console.debug(`Repository ID (cwdHash): ${cwdHash}`)
-
-			const newTracker = new CheckpointTracker(globalStoragePath, taskId, workingDir, cwdHash)
-
-			// Check if this is a legacy task
-			newTracker.isLegacyCheckpoint = await detectLegacyCheckpoint(newTracker.globalStoragePath, newTracker.taskId)
-			if (newTracker.isLegacyCheckpoint) {
-				console.debug("Using legacy checkpoint path structure")
-				const gitPath = await getShadowGitPath(
-					newTracker.globalStoragePath,
-					newTracker.taskId,
-					newTracker.cwdHash,
-					newTracker.isLegacyCheckpoint,
-				)
-				await GitOperations.initShadowGit(gitPath, workingDir, newTracker.isLegacyCheckpoint)
-				await newTracker.gitOperations.switchToTaskBranch(newTracker.taskId, gitPath)
-				return newTracker
-			}
-
-			// Branch-per-task structure
-			const gitPath = await getShadowGitPath(
-				newTracker.globalStoragePath,
-				newTracker.taskId,
-				newTracker.cwdHash,
-				newTracker.isLegacyCheckpoint,
-			)
-			await GitOperations.initShadowGit(gitPath, workingDir, newTracker.isLegacyCheckpoint)
-			await newTracker.gitOperations.switchToTaskBranch(newTracker.taskId, gitPath)
+			const cwd = await CheckpointTracker.getWorkingDirectory()
+			const newTracker = new CheckpointTracker(provider, taskId, cwd)
+			await newTracker.initShadowGit()
 			return newTracker
 		} catch (error) {
 			console.error("Failed to create CheckpointTracker:", error)
@@ -146,110 +50,203 @@ class CheckpointTracker {
 		}
 	}

-	/**
-	 * Creates a new checkpoint commit in the shadow git repository.
-	 *
-	 * Key behaviors:
-	 * - Creates commit with checkpoint files in shadow git repo
-	 * - Handles both legacy and branch-per-task checkpoint structures
-	 * - For new tasks, switches to task-specific branch first
-	 * - Caches the created commit hash
-	 *
-	 * Commit structure:
-	 * - Legacy: Simple "checkpoint" message
-	 * - Branch-per-task: "checkpoint-{cwdHash}-{taskId}"
-	 * - Always allows empty commits
-	 *
-	 * Dependencies:
-	 * - Requires initialized shadow git (getShadowGitPath)
-	 * - For new checkpoints, requires task branch setup
-	 * - Uses addCheckpointFiles to stage changes
-	 *
-	 * @returns Promise<string | undefined> The created commit hash, or undefined if:
-	 * - Shadow git access fails
-	 * - Branch switch fails
-	 * - Staging files fails
-	 * - Commit creation fails
-	 * @throws Error if unable to:
-	 * - Access shadow git path
-	 * - Initialize simple-git
-	 * - Switch branches
-	 * - Stage or commit files
-	 */
-	public async commit(): Promise<string | undefined> {
-		try {
-			console.info(`Creating new checkpoint commit for task ${this.taskId}`)
-			const gitPath = await getShadowGitPath(this.globalStoragePath, this.taskId, this.cwdHash, this.isLegacyCheckpoint)
-			const git = simpleGit(path.dirname(gitPath))
+	private static async getWorkingDirectory(): Promise<string> {
+		const cwd = vscode.workspace.workspaceFolders?.map((folder) => folder.uri.fsPath).at(0)
+		if (!cwd) {
+			throw new Error("No workspace detected. Please open Cline in a workspace to use checkpoints.")
+		}
+		const homedir = os.homedir()
+		const desktopPath = path.join(homedir, "Desktop")
+		const documentsPath = path.join(homedir, "Documents")
+		const downloadsPath = path.join(homedir, "Downloads")

-			console.info(`Using shadow git at: ${gitPath}`)
+		switch (cwd) {
+			case homedir:
+				throw new Error("Cannot use checkpoints in home directory")
+			case desktopPath:
+				throw new Error("Cannot use checkpoints in Desktop directory")
+			case documentsPath:
+				throw new Error("Cannot use checkpoints in Documents directory")
+			case downloadsPath:
+				throw new Error("Cannot use checkpoints in Downloads directory")
+			default:
+				return cwd
+		}
+	}

-			// Disable nested git repos before any operations
-			await this.gitOperations.renameNestedGitRepos(true)
+	private async getShadowGitPath(): Promise<string> {
+		const globalStoragePath = this.providerRef.deref()?.context.globalStorageUri.fsPath
+		if (!globalStoragePath) {
+			throw new Error("Global storage uri is invalid")
+		}
+		const checkpointsDir = path.join(globalStoragePath, "tasks", this.taskId, "checkpoints")
+		await fs.mkdir(checkpointsDir, { recursive: true })
+		const gitPath = path.join(checkpointsDir, ".git")
+		return gitPath
+	}

+	public static async doesShadowGitExist(taskId: string, provider?: ClineProvider): Promise<boolean> {
+		const globalStoragePath = provider?.context.globalStorageUri.fsPath
+		if (!globalStoragePath) {
+			return false
+		}
+		const gitPath = path.join(globalStoragePath, "tasks", taskId, "checkpoints", ".git")
+		return await fileExistsAtPath(gitPath)
+	}
+
+	public async initShadowGit(): Promise<string> {
+		const gitPath = await this.getShadowGitPath()
+		if (await fileExistsAtPath(gitPath)) {
+			// Make sure it's the same cwd as the configured worktree
+			const worktree = await this.getShadowGitConfigWorkTree()
+			if (worktree !== this.cwd) {
+				throw new Error("Checkpoints can only be used in the original workspace: " + worktree)
+			}
+
+			return gitPath
+		} else {
+			const checkpointsDir = path.dirname(gitPath)
+			const git = simpleGit(checkpointsDir)
+			await git.init()
+
+			await git.addConfig("core.worktree", this.cwd) // sets the working tree to the current workspace
+
+			// Disable commit signing for shadow repo
+			await git.addConfig("commit.gpgSign", "false")
+
+			// Get LFS patterns from workspace if they exist
+			let lfsPatterns: string[] = []
 			try {
-				if (!this.isLegacyCheckpoint) {
-					await this.gitOperations.switchToTaskBranch(this.taskId, gitPath)
+				const attributesPath = path.join(this.cwd, ".gitattributes")
+				if (await fileExistsAtPath(attributesPath)) {
+					const attributesContent = await fs.readFile(attributesPath, "utf8")
+					lfsPatterns = attributesContent
+						.split("\n")
+						.filter((line) => line.includes("filter=lfs"))
+						.map((line) => line.split(" ")[0].trim())
 				}
-				await this.gitOperations.addCheckpointFiles(git, gitPath)
-
-				const commitMessage = this.isLegacyCheckpoint ? "checkpoint" : "checkpoint-" + this.cwdHash + "-" + this.taskId
-
-				console.info(
-					`Creating ${this.isLegacyCheckpoint ? "legacy" : "new"} checkpoint commit with message: ${commitMessage}`,
-				)
-				const result = await git.commit(commitMessage, {
-					"--allow-empty": null,
-				})
-				const commitHash = result.commit || ""
-				this.lastCheckpointHash = commitHash
-				console.warn(`Checkpoint commit created.`)
-				return commitHash
-			} finally {
-				// Always re-enable nested git repos
-				await this.gitOperations.renameNestedGitRepos(false)
+			} catch (error) {
+				console.warn("Failed to read .gitattributes:", error)
 			}
-		} catch (error) {
-			console.error("Failed to create checkpoint:", {
-				taskId: this.taskId,
-				error,
-				isLegacyCheckpoint: this.isLegacyCheckpoint,
-			})
-			throw new Error(`Failed to create checkpoint: ${error instanceof Error ? error.message : String(error)}`)
+
+			// Add basic excludes directly in git config, while respecting any .gitignore in the workspace
+			// .git/info/exclude is local to the shadow git repo, so it's not shared with the main repo - and won't conflict with user's .gitignore
+			// TODO: let user customize these
+			const excludesPath = path.join(gitPath, "info", "exclude")
+			await fs.mkdir(path.join(gitPath, "info"), { recursive: true })
+			await fs.writeFile(
+				excludesPath,
+				[
+					".git/", // ignore the user's .git
+					`.git${GIT_DISABLED_SUFFIX}/`, // ignore the disabled nested git repos
+					".DS_Store",
+					"*.log",
+					"node_modules/",
+					"__pycache__/",
+					"env/",
+					"venv/",
+					"target/dependency/",
+					"build/dependencies/",
+					"dist/",
+					"out/",
+					"bundle/",
+					"vendor/",
+					"tmp/",
+					"temp/",
+					"deps/",
+					"pkg/",
+					"Pods/",
+					// Media files
+					"*.jpg",
+					"*.jpeg",
+					"*.png",
+					"*.gif",
+					"*.bmp",
+					"*.ico",
+					// "*.svg",
+					"*.mp3",
+					"*.mp4",
+					"*.wav",
+					"*.avi",
+					"*.mov",
+					"*.wmv",
+					"*.webm",
+					"*.webp",
+					"*.m4a",
+					"*.flac",
+					// Build and dependency directories
+					"build/",
+					"bin/",
+					"obj/",
+					".gradle/",
+					".idea/",
+					".vscode/",
+					".vs/",
+					"coverage/",
+					".next/",
+					".nuxt/",
+					// Cache and temporary files
+					"*.cache",
+					"*.tmp",
+					"*.temp",
+					"*.swp",
+					"*.swo",
+					"*.pyc",
+					"*.pyo",
+					".pytest_cache/",
+					".eslintcache",
+					// Environment and config files
+					".env*",
+					"*.local",
+					"*.development",
+					"*.production",
+					// Large data files
+					"*.zip",
+					"*.tar",
+					"*.gz",
+					"*.rar",
+					"*.7z",
+					"*.iso",
+					"*.bin",
+					"*.exe",
+					"*.dll",
+					"*.so",
+					"*.dylib",
+					// Database files
+					"*.sqlite",
+					"*.db",
+					"*.sql",
+					// Log files
+					"*.logs",
+					"*.error",
+					"npm-debug.log*",
+					"yarn-debug.log*",
+					"yarn-error.log*",
+					...lfsPatterns,
+				].join("\n"),
+			)
+
+			// Set up git identity (git throws an error if user.name or user.email is not set)
+			await git.addConfig("user.name", "Cline Checkpoint")
+			await git.addConfig("user.email", "noreply@example.com")
+
+			await this.addAllFiles(git)
+			// Initial commit (--allow-empty ensures it works even with no files)
+			await git.commit("initial commit", { "--allow-empty": null })
+
+			return gitPath
 		}
 	}

-	/**
-	 * Retrieves the worktree path from the shadow git configuration.
-	 * The worktree path indicates where the shadow git repository is tracking files,
-	 * which should match the current workspace directory.
-	 *
-	 * Key behaviors:
-	 * - Caches result in lastRetrievedShadowGitConfigWorkTree to avoid repeated reads
-	 * - Returns cached value if available
-	 * - Reads git config if no cached value exists
-	 * - Handles both legacy and new checkpoint structures
-	 *
-	 * Configuration read:
-	 * - Uses simple-git to read core.worktree config
-	 * - Operates on shadow git at path from getShadowGitPath()
-	 *
-	 * @returns Promise<string | undefined> The configured worktree path, or undefined if:
-	 * - Shadow git repository doesn't exist
-	 * - Config read fails
-	 * - No worktree is configured
-	 * @throws Error if unable to:
-	 * - Access shadow git path
-	 * - Initialize simple-git
-	 * - Read git configuration
-	 */
 	public async getShadowGitConfigWorkTree(): Promise<string | undefined> {
 		if (this.lastRetrievedShadowGitConfigWorkTree) {
 			return this.lastRetrievedShadowGitConfigWorkTree
 		}
 		try {
-			const gitPath = await getShadowGitPath(this.globalStoragePath, this.taskId, this.cwdHash, this.isLegacyCheckpoint)
-			this.lastRetrievedShadowGitConfigWorkTree = await this.gitOperations.getShadowGitConfigWorkTree(gitPath)
+			const gitPath = await this.getShadowGitPath()
+			const git = simpleGit(path.dirname(gitPath))
+			const worktree = await git.getConfig("core.worktree")
+			this.lastRetrievedShadowGitConfigWorkTree = worktree.value || undefined
 			return this.lastRetrievedShadowGitConfigWorkTree
 		} catch (error) {
 			console.error("Failed to get shadow git config worktree:", error)
@@ -257,32 +254,36 @@ class CheckpointTracker {
 		}
 	}

-	/**
-	 * Resets the shadow git repository's HEAD to a specific checkpoint commit.
-	 * This will discard all changes after the target commit and restore the
-	 * working directory to that checkpoint's state.
-	 *
-	 * Dependencies:
-	 * - Requires initialized shadow git (getShadowGitPath)
-	 * - For new checkpoints, requires task branch setup
-	 * - Must be called with a valid commit hash from this task's history
-	 *
-	 * @param commitHash - The hash of the checkpoint commit to reset to
-	 * @returns Promise<void> Resolves when reset is complete
-	 * @throws Error if unable to:
-	 * - Access shadow git path
-	 * - Initialize simple-git
-	 * - Switch to task branch
-	 * - Reset to target commit
-	 */
+	public async commit(): Promise<string | undefined> {
+		try {
+			const gitPath = await this.getShadowGitPath()
+			const git = simpleGit(path.dirname(gitPath))
+			await this.addAllFiles(git)
+			const result = await git.commit("checkpoint", {
+				"--allow-empty": null,
+			})
+			const commitHash = result.commit || ""
+			this.lastCheckpointHash = commitHash
+			return commitHash
+		} catch (error) {
+			console.error("Failed to create checkpoint:", error)
+			return undefined
+		}
+	}
+
 	public async resetHead(commitHash: string): Promise<void> {
-		console.info(`Resetting to checkpoint: ${commitHash}`)
-		const gitPath = await getShadowGitPath(this.globalStoragePath, this.taskId, this.cwdHash, this.isLegacyCheckpoint)
+		const gitPath = await this.getShadowGitPath()
 		const git = simpleGit(path.dirname(gitPath))
-		console.debug(`Using shadow git at: ${gitPath}`)
-		await this.gitOperations.switchToTaskBranch(this.taskId, gitPath)
+
+		// Clean working directory and force reset
+		// This ensures that the operation will succeed regardless of:
+		// - Untracked files in the workspace
+		// - Staged changes
+		// - Unstaged changes
+		// - Partial commits
+		// - Merge conflicts
+		await git.clean("f", ["-d", "-f"]) // Remove untracked files and directories
 		await git.reset(["--hard", commitHash]) // Hard reset to target commit
-		console.debug(`Successfully reset to checkpoint: ${commitHash}`)
 	}

 	/**
@@ -309,145 +310,111 @@ class CheckpointTracker {
 			after: string
 		}>
 	> {
-		const gitPath = await getShadowGitPath(this.globalStoragePath, this.taskId, this.cwdHash, this.isLegacyCheckpoint)
+		const gitPath = await this.getShadowGitPath()
 		const git = simpleGit(path.dirname(gitPath))

-		if (!this.isLegacyCheckpoint) {
-			await this.gitOperations.switchToTaskBranch(this.taskId, gitPath)
-		}
-
-		console.info(`Getting diff between commits: ${lhsHash || "initial"} -> ${rhsHash || "working directory"}`)
-
 		// If lhsHash is missing, use the initial commit of the repo
 		let baseHash = lhsHash
 		if (!baseHash) {
 			const rootCommit = await git.raw(["rev-list", "--max-parents=0", "HEAD"])
 			baseHash = rootCommit.trim()
-			console.debug(`Using root commit as base: ${baseHash}`)
 		}

 		// Stage all changes so that untracked files appear in diff summary
-		await this.gitOperations.addCheckpointFiles(git, gitPath)
+		await this.addAllFiles(git)

 		const diffSummary = rhsHash ? await git.diffSummary([`${baseHash}..${rhsHash}`]) : await git.diffSummary([baseHash])
-		console.info(`Found ${diffSummary.files.length} changed files`)

 		// For each changed file, gather before/after content
 		const result = []
 		const cwdPath = (await this.getShadowGitConfigWorkTree()) || this.cwd || ""
-		const files = diffSummary.files.map((f) => f.file)
-		const batchSize = 50

-		// Get list of files that exist in base commit
-		const existingFiles = await this.getExistingFiles(git, baseHash, files)
+		for (const file of diffSummary.files) {
+			const filePath = file.file
+			const absolutePath = path.join(cwdPath, filePath)

-		// Process files in batches
-		for (let i = 0; i < files.length; i += batchSize) {
-			const batch = files.slice(i, i + batchSize)
-
-			// Split batch into existing and new files
-			const existingBatch = batch.filter((file) => existingFiles.has(file))
-			const newBatch = batch.filter((file) => !existingFiles.has(file))
-
-			// Get before contents for existing files
-			let beforeContents: string[] = new Array(batch.length).fill("")
-			if (existingBatch.length > 0) {
-				await git.addConfig("core.quotePath", "false")
-				await git.addConfig("core.precomposeunicode", "true")
-				const args = ["show", "--format="]
-				existingBatch.forEach((file) => {
-					args.push(`${baseHash}:${file}`)
-				})
-				const beforeResult = await git.raw(args)
-				const existingContents = beforeResult.split("\n\0\n")
-				// Map contents back to original batch positions
-				existingBatch.forEach((file, index) => {
-					const batchIndex = batch.indexOf(file)
-					if (batchIndex !== -1) {
-						beforeContents[batchIndex] = existingContents[index] || ""
-					}
-				})
+			let beforeContent = ""
+			try {
+				beforeContent = await git.show([`${baseHash}:${filePath}`])
+			} catch (_) {
+				// file didn't exist in older commit => remains empty
 			}

-			// Get after contents
-			let afterContents: string[] = []
+			let afterContent = ""
 			if (rhsHash) {
-				// Split after files into existing and new in target commit
-				const afterExistingFiles = await this.getExistingFiles(git, rhsHash, batch)
-				const afterExistingBatch = batch.filter((file) => afterExistingFiles.has(file))
-
-				if (afterExistingBatch.length > 0) {
-					const args = ["show", "--format="]
-					afterExistingBatch.forEach((file) => {
-						args.push(`${rhsHash}:${file}`)
-					})
-					const afterResult = await git.raw(args)
-					const existingContents = afterResult.split("\n\0\n")
-					afterContents = new Array(batch.length).fill("")
-					afterExistingBatch.forEach((file, index) => {
-						const batchIndex = batch.indexOf(file)
-						if (batchIndex !== -1) {
-							afterContents[batchIndex] = existingContents[index] || ""
-						}
-					})
+				// if user provided a newer commit, use git.show at that commit
+				try {
+					afterContent = await git.show([`${rhsHash}:${filePath}`])
+				} catch (_) {
+					// file didn't exist in newer commit => remains empty
 				}
 			} else {
-				// Read from disk for working directory changes
-				afterContents = await Promise.all(
-					batch.map(async (filePath) => {
-						try {
-							return await fs.readFile(path.join(cwdPath, filePath), "utf8")
-						} catch (_) {
-							return ""
-						}
-					}),
-				)
+				// otherwise, read from disk (includes uncommitted changes)
+				try {
+					afterContent = await fs.readFile(absolutePath, "utf8")
+				} catch (_) {
+					// file might be deleted => remains empty
+				}
 			}

-			// Add results for this batch
-			for (let j = 0; j < batch.length; j++) {
-				const filePath = batch[j]
-				const absolutePath = path.join(cwdPath, filePath)
-				result.push({
-					relativePath: filePath,
-					absolutePath,
-					before: beforeContents[j] || "",
-					after: afterContents[j] || "",
-				})
-			}
+			result.push({
+				relativePath: filePath,
+				absolutePath,
+				before: beforeContent,
+				after: afterContent,
+			})
 		}
+
 		return result
 	}

-	/**
-	 * Deletes all checkpoint data for a given task.
-	 * Handles both legacy checkpoints and branch-per-task checkpoints.
-	 *
-	 * @param taskId - The ID of the task whose checkpoints should be deleted
-	 * @param historyItem - The history item containing the shadow git config for this task
-	 * @param globalStoragePath - the globalStorage path
-	 * @throws Error if deletion fails
-	 */
-	public static async deleteCheckpoints(taskId: string, historyItem: HistoryItem, globalStoragePath: string): Promise<void> {
-		if (!globalStoragePath) {
-			throw new Error("Global storage uri is invalid")
+	private async addAllFiles(git: SimpleGit) {
+		await this.renameNestedGitRepos(true)
+		try {
+			await git.add(".")
+		} catch (error) {
+			console.error("Failed to add files to git:", error)
+		} finally {
+			await this.renameNestedGitRepos(false)
 		}
-		await GitOperations.deleteTaskBranchStatic(taskId, historyItem, globalStoragePath)
 	}

-	/**
-	 * Helper function to get a set of files that exist in a given commit
-	 */
-	private async getExistingFiles(git: SimpleGit, commitHash: string, files: string[]): Promise<Set<string>> {
-		try {
-			const result = await git.raw(["ls-tree", "-r", "--name-only", commitHash])
-			const existingFiles = new Set<string>(result.split("\n"))
-			return existingFiles
-		} catch (error) {
-			console.error("Error getting existing files:", error)
-			return new Set()
+	// Since we use git to track checkpoints, we need to temporarily disable nested git repos to work around git's requirement of using submodules for nested repos.
+	private async renameNestedGitRepos(disable: boolean) {
+		// Find all .git directories that are not at the root level
+		const gitPaths = await globby("**/.git" + (disable ? "" : GIT_DISABLED_SUFFIX), {
+			cwd: this.cwd,
+			onlyDirectories: true,
+			ignore: [".git"], // Ignore root level .git
+			dot: true,
+			markDirectories: false,
+		})
+
+		// For each nested .git directory, rename it based on operation
+		for (const gitPath of gitPaths) {
+			const fullPath = path.join(this.cwd, gitPath)
+			let newPath: string
+			if (disable) {
+				newPath = fullPath + GIT_DISABLED_SUFFIX
+			} else {
+				newPath = fullPath.endsWith(GIT_DISABLED_SUFFIX) ? fullPath.slice(0, -GIT_DISABLED_SUFFIX.length) : fullPath
+			}
+
+			try {
+				await fs.rename(fullPath, newPath)
+				console.log(`CheckpointTracker ${disable ? "disabled" : "enabled"} nested git repo ${gitPath}`)
+			} catch (error) {
+				console.error(`CheckpointTracker failed to ${disable ? "disable" : "enable"} nested git repo ${gitPath}:`, error)
+			}
 		}
 	}
+
+	public dispose() {
+		this.disposables.forEach((d) => d.dispose())
+		this.disposables = []
+	}
 }

+const GIT_DISABLED_SUFFIX = "_disabled"
+
 export default CheckpointTracker
@@ -47,10 +47,7 @@ export async function downloadTask(dateTs: number, conversationHistory: Anthropi
 	}
 }

-export function formatContentBlockToMarkdown(
-	block: Anthropic.ContentBlockParam,
-	// messages: Anthropic.MessageParam[]
-): string {
+export function formatContentBlockToMarkdown(block: Anthropic.ContentBlockParam): string {
 	switch (block.type) {
 		case "text":
 			return block.text
@@ -69,32 +66,16 @@ export function formatContentBlockToMarkdown(
 			}
 			return `[Tool Use: ${block.name}]\n${input}`
 		case "tool_result":
-			// For now we're not doing tool name lookup since we don't use tools anymore
-			// const toolName = findToolName(block.tool_use_id, messages)
-			const toolName = "Tool"
 			if (typeof block.content === "string") {
-				return `[${toolName}${block.is_error ? " (Error)" : ""}]\n${block.content}`
+				return `[Tool${block.is_error ? " (Error)" : ""}]\n${block.content}`
 			} else if (Array.isArray(block.content)) {
-				return `[${toolName}${block.is_error ? " (Error)" : ""}]\n${block.content
+				return `[Tool${block.is_error ? " (Error)" : ""}]\n${block.content
 					.map((contentBlock) => formatContentBlockToMarkdown(contentBlock))
 					.join("\n")}`
 			} else {
-				return `[${toolName}${block.is_error ? " (Error)" : ""}]`
+				return `[Tool${block.is_error ? " (Error)" : ""}]`
 			}
 		default:
 			return "[Unexpected content type]"
 	}
 }
-
-export function findToolName(toolCallId: string, messages: Anthropic.MessageParam[]): string {
-	for (const message of messages) {
-		if (Array.isArray(message.content)) {
-			for (const block of message.content) {
-				if (block.type === "tool_use" && block.id === toolCallId) {
-					return block.name
-				}
-			}
-		}
-	}
-	return "Unknown Tool"
-}
@@ -571,7 +571,6 @@ export class McpHub {
 			timeout = secondsToMs(parsedConfig.timeout)
 		} catch (error) {
 			console.error(`Failed to parse timeout configuration for server ${serverName}: ${error}`)
-			// Continue with default timeout
 		}

 		return await connection.client.request(
@@ -611,7 +610,6 @@ export class McpHub {
 				autoApprove.splice(toolIndex, 1)
 			}

-			// Write updated config back to file
 			await fs.writeFile(settingsPath, JSON.stringify(config, null, 2))

 			// Update the tools list to reflect the change
@@ -670,19 +668,14 @@ export class McpHub {
 				throw new Error(`Server "${serverName}" not found in settings`)
 			}

-			// Update the timeout in the config
 			config.mcpServers[serverName] = {
 				...config.mcpServers[serverName],
 				timeout,
 			}

-			// Write updated config back to file
 			await fs.writeFile(settingsPath, JSON.stringify(config, null, 2))

-			// Update server connections to apply the new timeout
 			await this.updateServerConnections(config.mcpServers)
-
-			vscode.window.showInformationMessage(`Updated timeout to ${timeout} seconds`)
 		} catch (error) {
 			console.error("Failed to update server timeout:", error)
 			if (error instanceof Error) {
@@ -22,7 +22,6 @@ export interface ExtensionMessage {
 		| "invoke"
 		| "partialMessage"
 		| "openRouterModels"
-		| "requestyModels"
 		| "openAiModels"
 		| "mcpServers"
 		| "relinquishControl"
@@ -52,7 +51,6 @@ export interface ExtensionMessage {
 	filePaths?: string[]
 	partialMessage?: ClineMessage
 	openRouterModels?: Record<string, ModelInfo>
-	requestyModels?: Record<string, ModelInfo>
 	openAiModels?: string[]
 	mcpServers?: McpServer[]
 	mcpMarketplaceCatalog?: McpMarketplaceCatalog
@@ -27,7 +27,6 @@ export interface WebviewMessage {
 		| "openMention"
 		| "cancelTask"
 		| "refreshOpenRouterModels"
-		| "refreshRequestyModels"
 		| "refreshOpenAiModels"
 		| "openMcpSettings"
 		| "restartMcpServer"
@@ -70,8 +69,7 @@ export interface WebviewMessage {
 	chatSettings?: ChatSettings
 	chatContent?: ChatContent
 	mcpId?: string
-	timeout?: number // For updateMcpTimeout
-
+	timeout?: number
 	// For toggleToolAutoApprove
 	serverName?: string
 	toolName?: string
@@ -15,6 +15,7 @@ export type ApiProvider =
 	| "mistral"
 	| "vscode-lm"
 	| "litellm"
+	| "asksage"
 	| "xai"

 export interface ApiHandlerOptions {
@@ -32,6 +33,7 @@ export interface ApiHandlerOptions {
 	awsSessionToken?: string
 	awsRegion?: string
 	awsUseCrossRegionInference?: boolean
+	awsBedrockUsePromptCache?: boolean
 	awsUseProfile?: boolean
 	awsProfile?: string
 	vertexProjectId?: string
@@ -42,6 +44,7 @@ export interface ApiHandlerOptions {
 	openAiModelInfo?: ModelInfo
 	ollamaModelId?: string
 	ollamaBaseUrl?: string
+	ollamaApiOptionsCtxNum?: string
 	lmStudioModelId?: string
 	lmStudioBaseUrl?: string
 	geminiApiKey?: string
@@ -49,7 +52,6 @@ export interface ApiHandlerOptions {
 	deepSeekApiKey?: string
 	requestyApiKey?: string
 	requestyModelId?: string
-	requestyModelInfo?: ModelInfo
 	togetherApiKey?: string
 	togetherModelId?: string
 	qwenApiKey?: string
@@ -58,7 +60,10 @@ export interface ApiHandlerOptions {
 	vsCodeLmModelSelector?: any
 	o3MiniReasoningEffort?: string
 	qwenApiLine?: string
+	asksageApiUrl?: string
+	asksageApiKey?: string
 	xaiApiKey?: string
+	thinkingBudgetTokens?: number
 }

 export type ApiConfiguration = ApiHandlerOptions & {
@@ -160,17 +165,21 @@ export const bedrockModels = {
 		contextWindow: 200_000,
 		supportsImages: true,
 		supportsComputerUse: true,
-		supportsPromptCache: false,
+		supportsPromptCache: true,
 		inputPrice: 3.0,
 		outputPrice: 15.0,
+		cacheWritesPrice: 3.75,
+		cacheReadsPrice: 0.3,
 	},
 	"anthropic.claude-3-5-haiku-20241022-v1:0": {
 		maxTokens: 8192,
 		contextWindow: 200_000,
 		supportsImages: false,
-		supportsPromptCache: false,
+		supportsPromptCache: true,
 		inputPrice: 1.0,
 		outputPrice: 5.0,
+		cacheWritesPrice: 1.0,
+		cacheReadsPrice: 0.08,
 	},
 	"anthropic.claude-3-5-sonnet-20240620-v1:0": {
 		maxTokens: 8192,
@@ -288,6 +297,78 @@ export const vertexModels = {
 		cacheWritesPrice: 0.3,
 		cacheReadsPrice: 0.03,
 	},
+	"gemini-2.0-flash-001": {
+		maxTokens: 8192,
+		contextWindow: 1_048_576,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0.1,
+		outputPrice: 0.4,
+	},
+	"gemini-2.0-flash-thinking-exp-1219": {
+		maxTokens: 8192,
+		contextWindow: 32_767,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+	"gemini-2.0-flash-exp": {
+		maxTokens: 8192,
+		contextWindow: 1_048_576,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+	"gemini-exp-1206": {
+		maxTokens: 8192,
+		contextWindow: 2_097_152,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+	"gemini-1.5-flash-002": {
+		maxTokens: 8192,
+		contextWindow: 1_048_576,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+	"gemini-1.5-flash-exp-0827": {
+		maxTokens: 8192,
+		contextWindow: 1_048_576,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+	"gemini-1.5-flash-8b-exp-0827": {
+		maxTokens: 8192,
+		contextWindow: 1_048_576,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+	"gemini-1.5-pro-002": {
+		maxTokens: 8192,
+		contextWindow: 2_097_152,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+	"gemini-1.5-pro-exp-0827": {
+		maxTokens: 8192,
+		contextWindow: 2_097_152,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
 } as const satisfies Record<string, ModelInfo>

 export const openAiModelInfoSaneDefaults: ModelInfo = {
@@ -411,9 +492,10 @@ export const openAiNativeModels = {
 		maxTokens: 100_000,
 		contextWindow: 200_000,
 		supportsImages: false,
-		supportsPromptCache: false,
+		supportsPromptCache: true,
 		inputPrice: 1.1,
 		outputPrice: 4.4,
+		cacheReadsPrice: 0.55,
 	},
 	// don't support tool use yet
 	o1: {
@@ -423,38 +505,43 @@ export const openAiNativeModels = {
 		supportsPromptCache: false,
 		inputPrice: 15,
 		outputPrice: 60,
+		cacheReadsPrice: 7.5,
 	},
 	"o1-preview": {
 		maxTokens: 32_768,
 		contextWindow: 128_000,
 		supportsImages: true,
-		supportsPromptCache: false,
+		supportsPromptCache: true,
 		inputPrice: 15,
 		outputPrice: 60,
+		cacheReadsPrice: 7.5,
 	},
 	"o1-mini": {
 		maxTokens: 65_536,
 		contextWindow: 128_000,
 		supportsImages: true,
-		supportsPromptCache: false,
+		supportsPromptCache: true,
 		inputPrice: 1.1,
 		outputPrice: 4.4,
+		cacheReadsPrice: 0.55,
 	},
 	"gpt-4o": {
 		maxTokens: 4_096,
 		contextWindow: 128_000,
 		supportsImages: true,
-		supportsPromptCache: false,
+		supportsPromptCache: true,
 		inputPrice: 2.5,
 		outputPrice: 10,
+		cacheReadsPrice: 1.25,
 	},
 	"gpt-4o-mini": {
 		maxTokens: 16_384,
 		contextWindow: 128_000,
 		supportsImages: true,
-		supportsPromptCache: false,
+		supportsPromptCache: true,
 		inputPrice: 0.15,
 		outputPrice: 0.6,
+		cacheReadsPrice: 0.075,
 	},
 	"gpt-4.5-preview": {
 		maxTokens: 16_384,
@@ -480,8 +567,8 @@ export const deepSeekModels = {
 		maxTokens: 8_000,
 		contextWindow: 64_000,
 		supportsImages: false,
-		supportsPromptCache: true, // supports context caching, but not in the way anthropic does it (deepseek reports input tokens and reads/writes in the same usage report) FIXME: we need to show users cache stats how deepseek does it
-		inputPrice: 0, // technically there is no input price, it's all either a cache hit or miss (ApiOptions will not show this)
+		supportsPromptCache: true,
+		inputPrice: 0.27,
 		outputPrice: 1.1,
 		cacheWritesPrice: 0.27,
 		cacheReadsPrice: 0.07,
@@ -490,8 +577,8 @@ export const deepSeekModels = {
 		maxTokens: 8_000,
 		contextWindow: 64_000,
 		supportsImages: false,
-		supportsPromptCache: true, // supports context caching, but not in the way anthropic does it (deepseek reports input tokens and reads/writes in the same usage report) FIXME: we need to show users cache stats how deepseek does it
-		inputPrice: 0, // technically there is no input price, it's all either a cache hit or miss (ApiOptions will not show this)
+		supportsPromptCache: true,
+		inputPrice: 0.55,
 		outputPrice: 2.19,
 		cacheWritesPrice: 0.55,
 		cacheReadsPrice: 0.14,
@@ -797,20 +884,36 @@ export const liteLlmModelInfoSaneDefaults: ModelInfo = {
 	outputPrice: 0,
 }

-// Requesty
-// https://requesty.ai/models
-export const requestyDefaultModelId = "anthropic/claude-3-5-sonnet-latest"
-export const requestyDefaultModelInfo: ModelInfo = {
-	maxTokens: 8192,
-	contextWindow: 200_000,
-	supportsImages: true,
-	supportsComputerUse: false,
-	supportsPromptCache: true,
-	inputPrice: 3.0,
-	outputPrice: 15.0,
-	cacheWritesPrice: 3.75,
-	cacheReadsPrice: 0.3,
-	description: "Anthropic's most intelligent model. Highest level of intelligence and capability.",
+// AskSage Models
+// https://docs.asksage.ai/
+export type AskSageModelId = keyof typeof askSageModels
+export const askSageDefaultModelId: AskSageModelId = "claude-35-sonnet"
+export const askSageDefaultURL: string = "https://api.asksage.ai/server"
+export const askSageModels = {
+	"gpt-4o": {
+		maxTokens: 4096,
+		contextWindow: 128000,
+		supportsImages: false,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+	"gpt-4o-gov": {
+		maxTokens: 4096,
+		contextWindow: 128000,
+		supportsImages: false,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+	"claude-35-sonnet": {
+		maxTokens: 8192,
+		contextWindow: 200000,
+		supportsImages: false,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
 }

 // X AI
@@ -1,6 +1,5 @@
 export const DEFAULT_MCP_TIMEOUT_SECONDS = 60 // matches Anthropic's default timeout in their MCP SDK
 export const MIN_MCP_TIMEOUT_SECONDS = 1
-
 export type McpMode = "full" | "server-use-only" | "off"

 export type McpServer = {
@@ -1,10 +1,10 @@
 import { describe, it } from "mocha"
 import "should"
-import { calculateApiCost } from "./cost"
+import { calculateApiCostAnthropic, calculateApiCostOpenAI } from "./cost"
 import { ModelInfo } from "../shared/api"

 describe("Cost Utilities", () => {
-	describe("calculateApiCost", () => {
+	describe("calculateApiCostAnthropic", () => {
 		it("should calculate basic input/output costs", () => {
 			const modelInfo: ModelInfo = {
 				supportsPromptCache: false,
@@ -12,7 +12,7 @@ describe("Cost Utilities", () => {
 				outputPrice: 15.0, // $15 per million tokens
 			}

-			const cost = calculateApiCost(modelInfo, 1000, 500)
+			const cost = calculateApiCostAnthropic(modelInfo, 1000, 500)
 			// Input: (3.0 / 1_000_000) * 1000 = 0.003
 			// Output: (15.0 / 1_000_000) * 500 = 0.0075
 			// Total: 0.003 + 0.0075 = 0.0105
@@ -25,7 +25,7 @@ describe("Cost Utilities", () => {
 				// No prices specified
 			}

-			const cost = calculateApiCost(modelInfo, 1000, 500)
+			const cost = calculateApiCostAnthropic(modelInfo, 1000, 500)
 			cost.should.equal(0)
 		})

@@ -42,7 +42,7 @@ describe("Cost Utilities", () => {
 				cacheReadsPrice: 0.3,
 			}

-			const cost = calculateApiCost(modelInfo, 2000, 1000, 1500, 500)
+			const cost = calculateApiCostAnthropic(modelInfo, 2000, 1000, 1500, 500)
 			// Cache writes: (3.75 / 1_000_000) * 1500 = 0.005625
 			// Cache reads: (0.3 / 1_000_000) * 500 = 0.00015
 			// Input: (3.0 / 1_000_000) * 2000 = 0.006
@@ -60,7 +60,68 @@ describe("Cost Utilities", () => {
 				cacheReadsPrice: 0.3,
 			}

-			const cost = calculateApiCost(modelInfo, 0, 0, 0, 0)
+			const cost = calculateApiCostAnthropic(modelInfo, 0, 0, 0, 0)
+			cost.should.equal(0)
+		})
+	})
+
+	describe("calculateApiCostOpenAI", () => {
+		it("should calculate basic input/output costs", () => {
+			const modelInfo: ModelInfo = {
+				supportsPromptCache: false,
+				inputPrice: 3.0, // $3 per million tokens
+				outputPrice: 15.0, // $15 per million tokens
+			}
+
+			const cost = calculateApiCostOpenAI(modelInfo, 1000, 500)
+			// Input: (3.0 / 1_000_000) * 1000 = 0.003
+			// Output: (15.0 / 1_000_000) * 500 = 0.0075
+			// Total: 0.003 + 0.0075 = 0.0105
+			cost.should.equal(0.0105)
+		})
+
+		it("should handle missing prices", () => {
+			const modelInfo: ModelInfo = {
+				supportsPromptCache: true,
+				// No prices specified
+			}
+
+			const cost = calculateApiCostOpenAI(modelInfo, 1000, 500)
+			cost.should.equal(0)
+		})
+
+		it("should use real model configuration (Claude 3.5 Sonnet)", () => {
+			const modelInfo: ModelInfo = {
+				maxTokens: 8192,
+				contextWindow: 200_000,
+				supportsImages: true,
+				supportsComputerUse: true,
+				supportsPromptCache: true,
+				inputPrice: 3.0,
+				outputPrice: 15.0,
+				cacheWritesPrice: 3.75,
+				cacheReadsPrice: 0.3,
+			}
+
+			const cost = calculateApiCostOpenAI(modelInfo, 2100, 1000, 1500, 500)
+			// Cache writes: (3.75 / 1_000_000) * 1500 = 0.005625
+			// Cache reads: (0.3 / 1_000_000) * 500 = 0.00015
+			// Input: (3.0 / 1_000_000) * (2100 - 1500 - 500) = 0.0003
+			// Output: (15.0 / 1_000_000) * 1000 = 0.015
+			// Total: 0.005625 + 0.00015 + 0.0003 + 0.015 = 0.021075
+			cost.should.equal(0.021075)
+		})
+
+		it("should handle zero token counts", () => {
+			const modelInfo: ModelInfo = {
+				supportsPromptCache: true,
+				inputPrice: 3.0,
+				outputPrice: 15.0,
+				cacheWritesPrice: 3.75,
+				cacheReadsPrice: 0.3,
+			}
+
+			const cost = calculateApiCostOpenAI(modelInfo, 0, 0, 0, 0)
 			cost.should.equal(0)
 		})
 	})
@@ -1,24 +1,49 @@
 import { ModelInfo } from "../shared/api"

-export function calculateApiCost(
+function calculateApiCostInternal(
+	modelInfo: ModelInfo,
+	inputTokens: number,
+	outputTokens: number,
+	cacheCreationInputTokens: number,
+	cacheReadInputTokens: number,
+): number {
+	const cacheWritesCost = ((modelInfo.cacheWritesPrice || 0) / 1_000_000) * cacheCreationInputTokens
+	const cacheReadsCost = ((modelInfo.cacheReadsPrice || 0) / 1_000_000) * cacheReadInputTokens
+	const baseInputCost = ((modelInfo.inputPrice || 0) / 1_000_000) * inputTokens
+	const outputCost = ((modelInfo.outputPrice || 0) / 1_000_000) * outputTokens
+	const totalCost = cacheWritesCost + cacheReadsCost + baseInputCost + outputCost
+	return totalCost
+}
+
+// For Anthropic compliant usage, the input tokens count does NOT include the cached tokens
+export function calculateApiCostAnthropic(
 	modelInfo: ModelInfo,
 	inputTokens: number,
 	outputTokens: number,
 	cacheCreationInputTokens?: number,
 	cacheReadInputTokens?: number,
 ): number {
-	const modelCacheWritesPrice = modelInfo.cacheWritesPrice
-	let cacheWritesCost = 0
-	if (cacheCreationInputTokens && modelCacheWritesPrice) {
-		cacheWritesCost = (modelCacheWritesPrice / 1_000_000) * cacheCreationInputTokens
-	}
-	const modelCacheReadsPrice = modelInfo.cacheReadsPrice
-	let cacheReadsCost = 0
-	if (cacheReadInputTokens && modelCacheReadsPrice) {
-		cacheReadsCost = (modelCacheReadsPrice / 1_000_000) * cacheReadInputTokens
-	}
-	const baseInputCost = ((modelInfo.inputPrice || 0) / 1_000_000) * inputTokens
-	const outputCost = ((modelInfo.outputPrice || 0) / 1_000_000) * outputTokens
-	const totalCost = cacheWritesCost + cacheReadsCost + baseInputCost + outputCost
-	return totalCost
+	const cacheCreationInputTokensNum = cacheCreationInputTokens || 0
+	const cacheReadInputTokensNum = cacheReadInputTokens || 0
+	return calculateApiCostInternal(modelInfo, inputTokens, outputTokens, cacheCreationInputTokensNum, cacheReadInputTokensNum)
+}
+
+// For OpenAI compliant usage, the input tokens count INCLUDES the cached tokens
+export function calculateApiCostOpenAI(
+	modelInfo: ModelInfo,
+	inputTokens: number,
+	outputTokens: number,
+	cacheCreationInputTokens?: number,
+	cacheReadInputTokens?: number,
+): number {
+	const cacheCreationInputTokensNum = cacheCreationInputTokens || 0
+	const cacheReadInputTokensNum = cacheReadInputTokens || 0
+	const nonCachedInputTokens = Math.max(0, inputTokens - cacheCreationInputTokensNum - cacheReadInputTokensNum)
+	return calculateApiCostInternal(
+		modelInfo,
+		nonCachedInputTokens,
+		outputTokens,
+		cacheCreationInputTokensNum,
+		cacheReadInputTokensNum,
+	)
 }
@@ -0,0 +1,36 @@
+import { anthropicModels } from "../shared/api"
+
+/**
+ * Validates the thinking budget token value according to the specified rules:
+ * - If disabled (0), return as is
+ * - If enabled but less than minimum (1024), set to minimum
+ * - If greater than or equal to max tokens, set to max tokens - 1
+ * - Otherwise, return the original value
+ *
+ * @param value The thinking budget token value to validate
+ * @param maxTokens The maximum tokens for the current model
+ * @returns The validated thinking budget token value
+ */
+export function validateThinkingBudget(
+	value: number,
+	maxTokens: number = anthropicModels["claude-3-7-sonnet-20250219"].maxTokens,
+): number {
+	// If disabled (0), return as is
+	if (value === 0) {
+		return 0
+	}
+
+	// If enabled but less than minimum, set to minimum
+	if (value > 0 && value < 1024) {
+		return 1024
+	}
+
+	// If greater than or equal to max allowed tokens (80% of max tokens), cap at that value
+	const maxAllowedTokens = Math.floor(maxTokens * 0.8)
+	if (value >= maxAllowedTokens) {
+		return maxAllowedTokens
+	}
+
+	// Otherwise, return the original value
+	return value
+}
@@ -19,8 +19,13 @@ export default tseslint.config(
 		},
 		rules: {
 			...reactHooks.configs.recommended.rules,
-			"react-refresh/only-export-components": ["warn", { allowConstantExport: true }],
+			// "react-refresh/only-export-components": ["warn", { allowConstantExport: true }],
 			"@typescript-eslint/no-unused-vars": "off",
+			"@typescript-eslint/no-explicit-any": "off",
+			"@typescript-eslint/no-empty-object-type": "off",
+			"no-case-declarations": "off",
+			"react-hooks/exhaustive-deps": "off",
+			"prefer-const": "off",
 		},
 	},
 )
@@ -9,7 +9,6 @@ import AccountView from "./components/account/AccountView"
 import { ExtensionStateContextProvider, useExtensionState } from "./context/ExtensionStateContext"
 import { vscode } from "./utils/vscode"
 import McpView from "./components/mcp/McpView"
-import posthog from "posthog-js"

 const AppContent = () => {
 	const { didHydrateState, showWelcome, shouldShowAnnouncement, telemetrySetting, vscMachineId } = useExtensionState()
@@ -61,14 +60,14 @@ const AppContent = () => {

 	useEvent("message", handleMessage)

-	useEffect(() => {
-		if (telemetrySetting === "enabled") {
-			posthog.identify(vscMachineId)
-			posthog.opt_in_capturing()
-		} else {
-			posthog.opt_out_capturing()
-		}
-	}, [telemetrySetting, vscMachineId])
+	// useEffect(() => {
+	// 	if (telemetrySetting === "enabled") {
+	// 		posthog.identify(vscMachineId)
+	// 		posthog.opt_in_capturing()
+	// 	} else {
+	// 		posthog.opt_out_capturing()
+	// 	}
+	// }, [telemetrySetting, vscMachineId])

 	useEffect(() => {
 		if (shouldShowAnnouncement) {
@@ -827,7 +827,7 @@ export const ChatRowContent = ({ message, isExpanded, onToggleExpand, lastModifi
 									{isExpanded ? (
 										<div style={{ marginTop: -3 }}>
 											<span style={{ fontWeight: "bold", display: "block", marginBottom: "4px" }}>
-												Reasoning
+												Thinking
 												<span
 													className="codicon codicon-chevron-down"
 													style={{
@@ -841,7 +841,7 @@ export const ChatRowContent = ({ message, isExpanded, onToggleExpand, lastModifi
 										</div>
 									) : (
 										<div style={{ display: "flex", alignItems: "center" }}>
-											<span style={{ fontWeight: "bold", marginRight: "4px" }}>Reasoning:</span>
+											<span style={{ fontWeight: "bold", marginRight: "4px" }}>Thinking:</span>
 											<span
 												style={{
 													whiteSpace: "nowrap",
@@ -214,7 +214,7 @@ const ChatTextArea = forwardRef<HTMLTextAreaElement, ChatTextAreaProps>(
 		},
 		ref,
 	) => {
-		const { filePaths, chatSettings, apiConfiguration, openRouterModels, requestyModels, platform } = useExtensionState()
+		const { filePaths, chatSettings, apiConfiguration, openRouterModels, platform } = useExtensionState()
 		const [isTextAreaFocused, setIsTextAreaFocused] = useState(false)
 		const [gitCommits, setGitCommits] = useState<any[]>([])

@@ -635,14 +635,14 @@ const ChatTextArea = forwardRef<HTMLTextAreaElement, ChatTextAreaProps>(
 		// Separate the API config submission logic
 		const submitApiConfig = useCallback(() => {
 			const apiValidationResult = validateApiConfiguration(apiConfiguration)
-			const modelIdValidationResult = validateModelId(apiConfiguration, openRouterModels, requestyModels)
+			const modelIdValidationResult = validateModelId(apiConfiguration, openRouterModels)

 			if (!apiValidationResult && !modelIdValidationResult) {
 				vscode.postMessage({ type: "apiConfiguration", apiConfiguration })
 			} else {
 				vscode.postMessage({ type: "getLatestState" })
 			}
-		}, [apiConfiguration, openRouterModels, requestyModels])
+		}, [apiConfiguration, openRouterModels])

 		const onModeToggle = useCallback(() => {
 			// if (textAreaDisabled) return
@@ -8,46 +8,46 @@ import {
 	VSCodeTextField,
 } from "@vscode/webview-ui-toolkit/react"
 import { Fragment, memo, useCallback, useEffect, useMemo, useState } from "react"
+import ThinkingBudgetSlider from "./ThinkingBudgetSlider"
 import { useEvent, useInterval } from "react-use"
+import styled from "styled-components"
+import * as vscodemodels from "vscode"
 import {
-	ApiConfiguration,
-	ApiProvider,
-	ModelInfo,
 	anthropicDefaultModelId,
 	anthropicModels,
+	ApiConfiguration,
+	ApiProvider,
 	azureOpenAiDefaultApiVersion,
 	bedrockDefaultModelId,
 	bedrockModels,
 	deepSeekDefaultModelId,
 	deepSeekModels,
-	qwenDefaultModelId,
-	qwenModels,
 	geminiDefaultModelId,
 	geminiModels,
 	mistralDefaultModelId,
 	mistralModels,
+	ModelInfo,
 	openAiModelInfoSaneDefaults,
 	openAiNativeDefaultModelId,
 	openAiNativeModels,
 	openRouterDefaultModelId,
 	openRouterDefaultModelInfo,
-	requestyDefaultModelId,
-	requestyDefaultModelInfo,
+	qwenDefaultModelId,
+	qwenModels,
 	vertexDefaultModelId,
 	vertexModels,
+	askSageModels,
+	askSageDefaultModelId,
+	askSageDefaultURL,
 	xaiDefaultModelId,
 	xaiModels,
 } from "../../../../src/shared/api"
 import { ExtensionMessage } from "../../../../src/shared/ExtensionMessage"
 import { useExtensionState } from "../../context/ExtensionStateContext"
 import { vscode } from "../../utils/vscode"
-import VSCodeButtonLink from "../common/VSCodeButtonLink"
-import OpenRouterModelPicker from "./OpenRouterModelPicker"
-import RequestyModelPicker from "./RequestyModelPicker"
-import ModelDescriptionMarkdown from "./ModelDescriptionMarkdown"
-import styled from "styled-components"
-import * as vscodemodels from "vscode"
 import { getAsVar, VSC_DESCRIPTION_FOREGROUND } from "../../utils/vscStyles"
+import VSCodeButtonLink from "../common/VSCodeButtonLink"
+import OpenRouterModelPicker, { ModelDescriptionMarkdown } from "./OpenRouterModelPicker"

 interface ApiOptionsProps {
 	showModelOptions: boolean
@@ -57,7 +57,7 @@ interface ApiOptionsProps {
 }

 // This is necessary to ensure dropdown opens downward, important for when this is used in popup
-const DROPDOWN_Z_INDEX = 1001 // Higher than the Requesty/OpenRouterModelPicker's and ModelSelectorTooltip's z-index
+const DROPDOWN_Z_INDEX = 1001 // Higher than the OpenRouterModelPicker's and ModelSelectorTooltip's z-index

 const DropdownContainer = styled.div<{ zIndex?: number }>`
 	position: relative;
@@ -201,10 +201,40 @@ const ApiOptions = ({ showModelOptions, apiErrorMessage, modelIdErrorMessage, is
 					<VSCodeOption value="lmstudio">LM Studio</VSCodeOption>
 					<VSCodeOption value="ollama">Ollama</VSCodeOption>
 					<VSCodeOption value="litellm">LiteLLM</VSCodeOption>
+					<VSCodeOption value="asksage">AskSage</VSCodeOption>
 					<VSCodeOption value="xai">X AI</VSCodeOption>
 				</VSCodeDropdown>
 			</DropdownContainer>

+			{selectedProvider === "asksage" && (
+				<div>
+					<VSCodeTextField
+						value={apiConfiguration?.asksageApiKey || ""}
+						style={{ width: "100%" }}
+						type="password"
+						onInput={handleInputChange("asksageApiKey")}
+						placeholder="Enter API Key...">
+						<span style={{ fontWeight: 500 }}>AskSage API Key</span>
+					</VSCodeTextField>
+					<p
+						style={{
+							fontSize: "12px",
+							marginTop: 3,
+							color: "var(--vscode-descriptionForeground)",
+						}}>
+						This key is stored locally and only used to make API requests from this extension.
+					</p>
+					<VSCodeTextField
+						value={apiConfiguration?.asksageApiUrl || askSageDefaultURL}
+						style={{ width: "100%" }}
+						type="url"
+						onInput={handleInputChange("asksageApiUrl")}
+						placeholder="Enter AskSage API URL...">
+						<span style={{ fontWeight: 500 }}>AskSage API URL</span>
+					</VSCodeTextField>
+				</div>
+			)}
+
 			{selectedProvider === "anthropic" && (
 				<div>
 					<VSCodeTextField
@@ -540,17 +570,35 @@ const ApiOptions = ({ showModelOptions, apiErrorMessage, modelIdErrorMessage, is
 							{/* <VSCodeOption value="us-gov-east-1">us-gov-east-1</VSCodeOption> */}
 						</VSCodeDropdown>
 					</DropdownContainer>
-					<VSCodeCheckbox
-						checked={apiConfiguration?.awsUseCrossRegionInference || false}
-						onChange={(e: any) => {
-							const isChecked = e.target.checked === true
-							setApiConfiguration({
-								...apiConfiguration,
-								awsUseCrossRegionInference: isChecked,
-							})
-						}}>
-						Use cross-region inference
-					</VSCodeCheckbox>
+					<div style={{ display: "flex", flexDirection: "column" }}>
+						<VSCodeCheckbox
+							checked={apiConfiguration?.awsUseCrossRegionInference || false}
+							onChange={(e: any) => {
+								const isChecked = e.target.checked === true
+								setApiConfiguration({
+									...apiConfiguration,
+									awsUseCrossRegionInference: isChecked,
+								})
+							}}>
+							Use cross-region inference
+						</VSCodeCheckbox>
+
+						{selectedModelInfo.supportsPromptCache && (
+							<>
+								<VSCodeCheckbox
+									checked={apiConfiguration?.awsBedrockUsePromptCache || false}
+									onChange={(e: any) => {
+										const isChecked = e.target.checked === true
+										setApiConfiguration({
+											...apiConfiguration,
+											awsBedrockUsePromptCache: isChecked,
+										})
+									}}>
+									Use prompt caching (Beta)
+								</VSCodeCheckbox>
+							</>
+						)}
+					</div>
 					<p
 						style={{
 							fontSize: "12px",
@@ -848,7 +896,24 @@ const ApiOptions = ({ showModelOptions, apiErrorMessage, modelIdErrorMessage, is
 						placeholder="Enter API Key...">
 						<span style={{ fontWeight: 500 }}>API Key</span>
 					</VSCodeTextField>
-					{!apiConfiguration?.requestyApiKey && <a href="https://app.requesty.ai/manage-api">Get API Key</a>}
+					<VSCodeTextField
+						value={apiConfiguration?.requestyModelId || ""}
+						style={{ width: "100%" }}
+						onInput={handleInputChange("requestyModelId")}
+						placeholder={"Enter Model ID..."}>
+						<span style={{ fontWeight: 500 }}>Model ID</span>
+					</VSCodeTextField>
+					<p
+						style={{
+							fontSize: "12px",
+							marginTop: 3,
+							color: "var(--vscode-descriptionForeground)",
+						}}>
+						<span style={{ color: "var(--vscode-errorForeground)" }}>
+							(<span style={{ fontWeight: 500 }}>Note:</span> Cline uses complex prompts and works best with Claude
+							models. Less capable models may not work as expected.)
+						</span>
+					</p>
 				</div>
 			)}

@@ -1068,6 +1133,13 @@ const ApiOptions = ({ showModelOptions, apiErrorMessage, modelIdErrorMessage, is
 						placeholder={"e.g. llama3.1"}>
 						<span style={{ fontWeight: 500 }}>Model ID</span>
 					</VSCodeTextField>
+					<VSCodeTextField
+						value={apiConfiguration?.ollamaApiOptionsCtxNum || "32768"}
+						style={{ width: "100%" }}
+						onInput={handleInputChange("ollamaApiOptionsCtxNum")}
+						placeholder={"e.g. 32768"}>
+						<span style={{ fontWeight: 500 }}>Model Context Window</span>
+					</VSCodeTextField>
 					{ollamaModels.length > 0 && (
 						<VSCodeRadioGroup
 							value={
@@ -1164,11 +1236,12 @@ const ApiOptions = ({ showModelOptions, apiErrorMessage, modelIdErrorMessage, is
 			)}

 			{selectedProvider !== "openrouter" &&
-				selectedProvider !== "requesty" &&
 				selectedProvider !== "openai" &&
 				selectedProvider !== "ollama" &&
 				selectedProvider !== "lmstudio" &&
 				selectedProvider !== "vscode-lm" &&
+				selectedProvider !== "litellm" &&
+				selectedProvider !== "requesty" &&
 				showModelOptions && (
 					<>
 						<DropdownContainer zIndex={DROPDOWN_Z_INDEX - 2} className="dropdown-container">
@@ -1183,9 +1256,16 @@ const ApiOptions = ({ showModelOptions, apiErrorMessage, modelIdErrorMessage, is
 							{selectedProvider === "deepseek" && createDropdown(deepSeekModels)}
 							{selectedProvider === "qwen" && createDropdown(qwenModels)}
 							{selectedProvider === "mistral" && createDropdown(mistralModels)}
+							{selectedProvider === "asksage" && createDropdown(askSageModels)}
 							{selectedProvider === "xai" && createDropdown(xaiModels)}
 						</DropdownContainer>

+						{((selectedProvider === "anthropic" && selectedModelId === "claude-3-7-sonnet-20250219") ||
+							(selectedProvider === "bedrock" && selectedModelId === "anthropic.claude-3-7-sonnet-20250219-v1:0") ||
+							(selectedProvider === "vertex" && selectedModelId === "claude-3-7-sonnet@20250219")) && (
+							<ThinkingBudgetSlider apiConfiguration={apiConfiguration} setApiConfiguration={setApiConfiguration} />
+						)}
+
 						<ModelInfoView
 							selectedModelId={selectedModelId}
 							modelInfo={selectedModelInfo}
@@ -1197,7 +1277,6 @@ const ApiOptions = ({ showModelOptions, apiErrorMessage, modelIdErrorMessage, is
 				)}

 			{selectedProvider === "openrouter" && showModelOptions && <OpenRouterModelPicker isPopup={isPopup} />}
-			{selectedProvider === "requesty" && showModelOptions && <RequestyModelPicker isPopup={isPopup} />}

 			{modelIdErrorMessage && (
 				<p
@@ -1395,18 +1474,14 @@ export function normalizeApiConfiguration(apiConfiguration?: ApiConfiguration):
 			return getProviderData(qwenModels, qwenDefaultModelId)
 		case "mistral":
 			return getProviderData(mistralModels, mistralDefaultModelId)
+		case "asksage":
+			return getProviderData(askSageModels, askSageDefaultModelId)
 		case "openrouter":
 			return {
 				selectedProvider: provider,
 				selectedModelId: apiConfiguration?.openRouterModelId || openRouterDefaultModelId,
 				selectedModelInfo: apiConfiguration?.openRouterModelInfo || openRouterDefaultModelInfo,
 			}
-		case "requesty":
-			return {
-				selectedProvider: provider,
-				selectedModelId: apiConfiguration?.requestyModelId || requestyDefaultModelId,
-				selectedModelInfo: apiConfiguration?.requestyModelInfo || requestyDefaultModelInfo,
-			}
 		case "openai":
 			return {
 				selectedProvider: provider,
@@ -1,6 +1,7 @@
 import { VSCodeLink, VSCodeTextField } from "@vscode/webview-ui-toolkit/react"
 import Fuse from "fuse.js"
-import React, { KeyboardEvent, useEffect, useMemo, useRef, useState } from "react"
+import React, { KeyboardEvent, memo, useEffect, useMemo, useRef, useState } from "react"
+import { useRemark } from "react-remark"
 import { useMount } from "react-use"
 import styled from "styled-components"
 import { openRouterDefaultModelId } from "../../../../src/shared/api"
@@ -8,6 +9,8 @@ import { useExtensionState } from "../../context/ExtensionStateContext"
 import { vscode } from "../../utils/vscode"
 import { highlight } from "../history/HistoryView"
 import { ModelInfoView, normalizeApiConfiguration } from "./ApiOptions"
+import { CODE_BLOCK_BG_COLOR } from "../common/CodeBlock"
+import ThinkingBudgetSlider from "./ThinkingBudgetSlider"

 export interface OpenRouterModelPickerProps {
 	isPopup?: boolean
@@ -133,6 +136,13 @@ const OpenRouterModelPicker: React.FC<OpenRouterModelPickerProps> = ({ isPopup }
 		}
 	}, [selectedIndex])

+	const showBudgetSlider = useMemo(() => {
+		return (
+			selectedModelId?.toLowerCase().includes("claude-3-7-sonnet") ||
+			selectedModelId?.toLowerCase().includes("claude-3.7-sonnet")
+		)
+	}, [selectedModelId])
+
 	return (
 		<div style={{ width: "100%" }}>
 			<style>
@@ -204,13 +214,18 @@ const OpenRouterModelPicker: React.FC<OpenRouterModelPickerProps> = ({ isPopup }
 			</div>

 			{hasInfo ? (
-				<ModelInfoView
-					selectedModelId={selectedModelId}
-					modelInfo={selectedModelInfo}
-					isDescriptionExpanded={isDescriptionExpanded}
-					setIsDescriptionExpanded={setIsDescriptionExpanded}
-					isPopup={isPopup}
-				/>
+				<>
+					{showBudgetSlider && (
+						<ThinkingBudgetSlider apiConfiguration={apiConfiguration} setApiConfiguration={setApiConfiguration} />
+					)}
+					<ModelInfoView
+						selectedModelId={selectedModelId}
+						modelInfo={selectedModelInfo}
+						isDescriptionExpanded={isDescriptionExpanded}
+						setIsDescriptionExpanded={setIsDescriptionExpanded}
+						isPopup={isPopup}
+					/>
+				</>
 			) : (
 				<p
 					style={{
@@ -274,3 +289,158 @@ const DropdownItem = styled.div<{ isSelected: boolean }>`
 		background-color: var(--vscode-list-activeSelectionBackground);
 	}
 `
+
+// Markdown
+
+const StyledMarkdown = styled.div`
+	font-family:
+		var(--vscode-font-family),
+		system-ui,
+		-apple-system,
+		BlinkMacSystemFont,
+		"Segoe UI",
+		Roboto,
+		Oxygen,
+		Ubuntu,
+		Cantarell,
+		"Open Sans",
+		"Helvetica Neue",
+		sans-serif;
+	font-size: 12px;
+	color: var(--vscode-descriptionForeground);
+
+	p,
+	li,
+	ol,
+	ul {
+		line-height: 1.25;
+		margin: 0;
+	}
+
+	ol,
+	ul {
+		padding-left: 1.5em;
+		margin-left: 0;
+	}
+
+	p {
+		white-space: pre-wrap;
+	}
+
+	a {
+		text-decoration: none;
+	}
+	a {
+		&:hover {
+			text-decoration: underline;
+		}
+	}
+`
+
+export const ModelDescriptionMarkdown = memo(
+	({
+		markdown,
+		key,
+		isExpanded,
+		setIsExpanded,
+		isPopup,
+	}: {
+		markdown?: string
+		key: string
+		isExpanded: boolean
+		setIsExpanded: (isExpanded: boolean) => void
+		isPopup?: boolean
+	}) => {
+		const [reactContent, setMarkdown] = useRemark()
+		// const [isExpanded, setIsExpanded] = useState(false)
+		const [showSeeMore, setShowSeeMore] = useState(false)
+		const textContainerRef = useRef<HTMLDivElement>(null)
+		const textRef = useRef<HTMLDivElement>(null)
+
+		useEffect(() => {
+			setMarkdown(markdown || "")
+		}, [markdown, setMarkdown])
+
+		useEffect(() => {
+			if (textRef.current && textContainerRef.current) {
+				const { scrollHeight } = textRef.current
+				const { clientHeight } = textContainerRef.current
+				const isOverflowing = scrollHeight > clientHeight
+				setShowSeeMore(isOverflowing)
+				// if (!isOverflowing) {
+				// 	setIsExpanded(false)
+				// }
+			}
+		}, [reactContent, setIsExpanded])
+
+		return (
+			<StyledMarkdown key={key} style={{ display: "inline-block", marginBottom: 0 }}>
+				<div
+					ref={textContainerRef}
+					style={{
+						overflowY: isExpanded ? "auto" : "hidden",
+						position: "relative",
+						wordBreak: "break-word",
+						overflowWrap: "anywhere",
+					}}>
+					<div
+						ref={textRef}
+						style={{
+							display: "-webkit-box",
+							WebkitLineClamp: isExpanded ? "unset" : 3,
+							WebkitBoxOrient: "vertical",
+							overflow: "hidden",
+							// whiteSpace: "pre-wrap",
+							// wordBreak: "break-word",
+							// overflowWrap: "anywhere",
+						}}>
+						{reactContent}
+					</div>
+					{!isExpanded && showSeeMore && (
+						<div
+							style={{
+								position: "absolute",
+								right: 0,
+								bottom: 0,
+								display: "flex",
+								alignItems: "center",
+							}}>
+							<div
+								style={{
+									width: 30,
+									height: "1.2em",
+									background: "linear-gradient(to right, transparent, var(--vscode-sideBar-background))",
+								}}
+							/>
+							<VSCodeLink
+								style={{
+									// cursor: "pointer",
+									// color: "var(--vscode-textLink-foreground)",
+									fontSize: "inherit",
+									paddingRight: 0,
+									paddingLeft: 3,
+									backgroundColor: isPopup ? CODE_BLOCK_BG_COLOR : "var(--vscode-sideBar-background)",
+								}}
+								onClick={() => setIsExpanded(true)}>
+								See more
+							</VSCodeLink>
+						</div>
+					)}
+				</div>
+				{/* {isExpanded && showSeeMore && (
+				<div
+					style={{
+						cursor: "pointer",
+						color: "var(--vscode-textLink-foreground)",
+						marginLeft: "auto",
+						textAlign: "right",
+						paddingRight: 2,
+					}}
+					onClick={() => setIsExpanded(false)}>
+					See less
+				</div>
+			)} */}
+			</StyledMarkdown>
+		)
+	},
+)
@@ -1,274 +0,0 @@
-import { VSCodeLink, VSCodeTextField } from "@vscode/webview-ui-toolkit/react"
-import Fuse from "fuse.js"
-import React, { KeyboardEvent, useEffect, useMemo, useRef, useState } from "react"
-import { useMount } from "react-use"
-import styled from "styled-components"
-import { requestyDefaultModelId } from "../../../../src/shared/api"
-import { useExtensionState } from "../../context/ExtensionStateContext"
-import { vscode } from "../../utils/vscode"
-import { highlight } from "../history/HistoryView"
-import { ModelInfoView, normalizeApiConfiguration } from "./ApiOptions"
-
-export interface RequestyModelPickerProps {
-	isPopup?: boolean
-}
-
-const RequestyModelPicker: React.FC<RequestyModelPickerProps> = ({ isPopup }) => {
-	const { apiConfiguration, setApiConfiguration, requestyModels } = useExtensionState()
-	const [searchTerm, setSearchTerm] = useState(apiConfiguration?.requestyModelId || requestyDefaultModelId)
-	const [isDropdownVisible, setIsDropdownVisible] = useState(false)
-	const [selectedIndex, setSelectedIndex] = useState(-1)
-	const dropdownRef = useRef<HTMLDivElement>(null)
-	const itemRefs = useRef<(HTMLDivElement | null)[]>([])
-	const [isDescriptionExpanded, setIsDescriptionExpanded] = useState(false)
-	const dropdownListRef = useRef<HTMLDivElement>(null)
-
-	const handleModelChange = (newModelId: string) => {
-		// could be setting invalid model id/undefined info but validation will catch it
-		setApiConfiguration({
-			...apiConfiguration,
-			...{
-				requestyModelId: newModelId,
-				requestyModelInfo: requestyModels[newModelId],
-			},
-		})
-		setSearchTerm(newModelId)
-	}
-
-	const { selectedModelId, selectedModelInfo } = useMemo(() => {
-		return normalizeApiConfiguration(apiConfiguration)
-	}, [apiConfiguration])
-
-	useMount(() => {
-		vscode.postMessage({ type: "refreshRequestyModels" })
-	})
-
-	useEffect(() => {
-		const handleClickOutside = (event: MouseEvent) => {
-			if (dropdownRef.current && !dropdownRef.current.contains(event.target as Node)) {
-				setIsDropdownVisible(false)
-			}
-		}
-
-		document.addEventListener("mousedown", handleClickOutside)
-		return () => {
-			document.removeEventListener("mousedown", handleClickOutside)
-		}
-	}, [])
-
-	const modelIds = useMemo(() => {
-		return Object.keys(requestyModels).sort((a, b) => a.localeCompare(b))
-	}, [requestyModels])
-
-	const searchableItems = useMemo(() => {
-		return modelIds.map((id) => ({
-			id,
-			html: id,
-		}))
-	}, [modelIds])
-
-	const fuse = useMemo(() => {
-		return new Fuse(searchableItems, {
-			keys: ["html"], // highlight function will update this
-			threshold: 0.6,
-			shouldSort: true,
-			isCaseSensitive: false,
-			ignoreLocation: false,
-			includeMatches: true,
-			minMatchCharLength: 1,
-		})
-	}, [searchableItems])
-
-	const modelSearchResults = useMemo(() => {
-		let results: { id: string; html: string }[] = searchTerm
-			? highlight(fuse.search(searchTerm), "model-item-highlight")
-			: searchableItems
-		return results
-	}, [searchableItems, searchTerm, fuse])
-
-	const handleKeyDown = (event: KeyboardEvent<HTMLInputElement>) => {
-		if (!isDropdownVisible) return
-
-		switch (event.key) {
-			case "ArrowDown":
-				event.preventDefault()
-				setSelectedIndex((prev) => (prev < modelSearchResults.length - 1 ? prev + 1 : prev))
-				break
-			case "ArrowUp":
-				event.preventDefault()
-				setSelectedIndex((prev) => (prev > 0 ? prev - 1 : prev))
-				break
-			case "Enter":
-				event.preventDefault()
-				if (selectedIndex >= 0 && selectedIndex < modelSearchResults.length) {
-					handleModelChange(modelSearchResults[selectedIndex].id)
-					setIsDropdownVisible(false)
-				}
-				break
-			case "Escape":
-				setIsDropdownVisible(false)
-				setSelectedIndex(-1)
-				break
-		}
-	}
-
-	const hasInfo = useMemo(() => {
-		return modelIds.some((id) => id.toLowerCase() === searchTerm.toLowerCase())
-	}, [modelIds, searchTerm])
-
-	useEffect(() => {
-		setSelectedIndex(-1)
-		if (dropdownListRef.current) {
-			dropdownListRef.current.scrollTop = 0
-		}
-	}, [searchTerm])
-
-	useEffect(() => {
-		if (selectedIndex >= 0 && itemRefs.current[selectedIndex]) {
-			itemRefs.current[selectedIndex]?.scrollIntoView({
-				block: "nearest",
-				behavior: "smooth",
-			})
-		}
-	}, [selectedIndex])
-
-	return (
-		<div style={{ width: "100%" }}>
-			<style>
-				{`
-				.model-item-highlight {
-					background-color: var(--vscode-editor-findMatchHighlightBackground);
-					color: inherit;
-				}
-				`}
-			</style>
-			<div style={{ display: "flex", flexDirection: "column" }}>
-				<label htmlFor="model-search">
-					<span style={{ fontWeight: 500 }}>Model</span>
-				</label>
-				<DropdownWrapper ref={dropdownRef}>
-					<VSCodeTextField
-						id="model-search"
-						placeholder="Search and select a model..."
-						value={searchTerm}
-						onInput={(e) => {
-							handleModelChange((e.target as HTMLInputElement)?.value?.toLowerCase())
-							setIsDropdownVisible(true)
-						}}
-						onFocus={() => setIsDropdownVisible(true)}
-						onKeyDown={handleKeyDown}
-						style={{
-							width: "100%",
-							zIndex: REQUESTY_MODEL_PICKER_Z_INDEX,
-							position: "relative",
-						}}>
-						{searchTerm && (
-							<div
-								className="input-icon-button codicon codicon-close"
-								aria-label="Clear search"
-								onClick={() => {
-									handleModelChange("")
-									setIsDropdownVisible(true)
-								}}
-								slot="end"
-								style={{
-									display: "flex",
-									justifyContent: "center",
-									alignItems: "center",
-									height: "100%",
-								}}
-							/>
-						)}
-					</VSCodeTextField>
-					{isDropdownVisible && (
-						<DropdownList ref={dropdownListRef}>
-							{modelSearchResults.map((item, index) => (
-								<DropdownItem
-									key={item.id}
-									ref={(el) => (itemRefs.current[index] = el)}
-									isSelected={index === selectedIndex}
-									onMouseEnter={() => setSelectedIndex(index)}
-									onClick={() => {
-										handleModelChange(item.id)
-										setIsDropdownVisible(false)
-									}}
-									dangerouslySetInnerHTML={{
-										__html: item.html,
-									}}
-								/>
-							))}
-						</DropdownList>
-					)}
-				</DropdownWrapper>
-			</div>
-
-			{hasInfo ? (
-				<ModelInfoView
-					selectedModelId={selectedModelId}
-					modelInfo={selectedModelInfo}
-					isDescriptionExpanded={isDescriptionExpanded}
-					setIsDescriptionExpanded={setIsDescriptionExpanded}
-					isPopup={isPopup}
-				/>
-			) : (
-				<p
-					style={{
-						fontSize: "12px",
-						marginTop: 0,
-						color: "var(--vscode-descriptionForeground)",
-					}}>
-					<>
-						The extension automatically fetches the latest list of models available on{" "}
-						<VSCodeLink style={{ display: "inline", fontSize: "inherit" }} href="https://app.requesty.ai/router/list">
-							Requesty.
-						</VSCodeLink>
-						If you're unsure which model to choose, Cline works best with{" "}
-						<VSCodeLink
-							style={{ display: "inline", fontSize: "inherit" }}
-							onClick={() => handleModelChange("anthropic/claude-3-5-sonnet-latest")}>
-							anthropic/claude-3-5-sonnet-latest.
-						</VSCodeLink>
-					</>
-				</p>
-			)}
-		</div>
-	)
-}
-
-export default RequestyModelPicker
-
-// Dropdown
-
-const DropdownWrapper = styled.div`
-	position: relative;
-	width: 100%;
-`
-
-export const REQUESTY_MODEL_PICKER_Z_INDEX = 1_000
-
-const DropdownList = styled.div`
-	position: absolute;
-	top: calc(100% - 3px);
-	left: 0;
-	width: calc(100% - 2px);
-	max-height: 200px;
-	overflow-y: auto;
-	background-color: var(--vscode-dropdown-background);
-	border: 1px solid var(--vscode-list-activeSelectionBackground);
-	z-index: ${REQUESTY_MODEL_PICKER_Z_INDEX - 1};
-	border-bottom-left-radius: 3px;
-	border-bottom-right-radius: 3px;
-`
-
-const DropdownItem = styled.div<{ isSelected: boolean }>`
-	padding: 5px 10px;
-	cursor: pointer;
-	word-break: break-all;
-	white-space: normal;
-
-	background-color: ${({ isSelected }) => (isSelected ? "var(--vscode-list-activeSelectionBackground)" : "inherit")};
-
-	&:hover {
-		background-color: var(--vscode-list-activeSelectionBackground);
-	}
-`
@@ -18,7 +18,6 @@ const SettingsView = ({ onDone }: SettingsViewProps) => {
 		customInstructions,
 		setCustomInstructions,
 		openRouterModels,
-		requestyModels,
 		telemetrySetting,
 		setTelemetrySetting,
 	} = useExtensionState()
@@ -27,7 +26,7 @@ const SettingsView = ({ onDone }: SettingsViewProps) => {

 	const handleSubmit = () => {
 		const apiValidationResult = validateApiConfiguration(apiConfiguration)
-		const modelIdValidationResult = validateModelId(apiConfiguration, openRouterModels, requestyModels)
+		const modelIdValidationResult = validateModelId(apiConfiguration, openRouterModels)

 		setApiErrorMessage(apiValidationResult)
 		setModelIdErrorMessage(modelIdValidationResult)
@@ -0,0 +1,151 @@
+import { memo, useState } from "react"
+import { anthropicModels, ApiConfiguration } from "../../../../src/shared/api"
+import { VSCodeCheckbox } from "@vscode/webview-ui-toolkit/react"
+import styled from "styled-components"
+
+// Constants
+const MIN_VALID_TOKENS = 1024
+const MAX_PERCENTAGE = 0.8
+const THUMB_SIZE = 16
+
+// Styled Components
+const Container = styled.div`
+	display: flex;
+	flex-direction: column;
+	gap: 10px;
+`
+
+const LabelContainer = styled.div`
+	display: flex;
+	justify-content: space-between;
+	flex-wrap: wrap;
+	gap: 12px;
+`
+
+const Label = styled.label`
+	font-weight: 500;
+	display: block;
+	margin-right: auto;
+`
+const Description = styled.p`
+	font-size: 12px;
+	margin-top: 0px;
+	margin-bottom: 0px;
+	color: var(--vscode-descriptionForeground);
+`
+
+const RangeInput = styled.input<{ $value: number; $min: number; $max: number }>`
+	width: 100%;
+	height: 8px;
+	appearance: none;
+	border-radius: 4px;
+	outline: none;
+	cursor: pointer;
+	margin: 5px 0 0;
+	padding: 0;
+	background: ${(props) => {
+		const percentage = ((props.$value - props.$min) / (props.$max - props.$min)) * 100
+		return `linear-gradient(to right, 
+			var(--vscode-progressBar-background) 0%,
+			var(--vscode-progressBar-background) ${percentage}%,
+			var(--vscode-scrollbarSlider-background) ${percentage}%,
+			var(--vscode-scrollbarSlider-background) 100%)`
+	}};
+
+	&::-webkit-slider-thumb {
+		appearance: none;
+		width: ${THUMB_SIZE}px;
+		height: ${THUMB_SIZE}px;
+		border-radius: 50%;
+		background: var(--vscode-foreground);
+		cursor: pointer;
+		border: 0px solid var(--vscode-progressBar-background);
+		box-shadow: 0 2px 4px rgba(0, 0, 0, 0.2);
+	}
+
+	&:focus {
+		outline: none;
+	}
+
+	&:focus::-webkit-slider-thumb,
+	&:hover::-webkit-slider-thumb {
+		border-color: var(--vscode-progressBar-background);
+		box-shadow: 0 2px 6px rgba(0, 0, 0, 0.3);
+	}
+
+	&:active::-webkit-slider-thumb {
+		outline: none;
+		border-color: var(--vscode-progressBar-background);
+	}
+`
+
+interface ThinkingBudgetSliderProps {
+	apiConfiguration: ApiConfiguration | undefined
+	setApiConfiguration: (apiConfiguration: ApiConfiguration) => void
+}
+
+const ThinkingBudgetSlider = ({ apiConfiguration, setApiConfiguration }: ThinkingBudgetSliderProps) => {
+	const maxTokens = anthropicModels["claude-3-7-sonnet-20250219"].maxTokens
+	const maxSliderValue = Math.floor(maxTokens * MAX_PERCENTAGE)
+	const isEnabled = (apiConfiguration?.thinkingBudgetTokens || 0) > 0
+
+	// Add local state for the slider value
+	const [localValue, setLocalValue] = useState(apiConfiguration?.thinkingBudgetTokens || 0)
+
+	const handleSliderChange = (event: React.ChangeEvent<HTMLInputElement>) => {
+		const value = parseInt(event.target.value, 10)
+		setLocalValue(value)
+	}
+
+	const handleSliderComplete = () => {
+		setApiConfiguration({
+			...apiConfiguration,
+			thinkingBudgetTokens: localValue,
+		})
+	}
+
+	const handleToggleChange = (event: any) => {
+		const isChecked = (event.target as HTMLInputElement).checked
+		const newValue = isChecked ? MIN_VALID_TOKENS : 0
+		setLocalValue(newValue)
+		setApiConfiguration({
+			...apiConfiguration,
+			thinkingBudgetTokens: newValue,
+		})
+	}
+
+	return (
+		<Container>
+			<VSCodeCheckbox checked={isEnabled} onChange={handleToggleChange}>
+				Enable extended thinking
+			</VSCodeCheckbox>
+
+			{isEnabled && (
+				<>
+					<LabelContainer>
+						<Label>
+							<strong>Budget:</strong> {localValue.toLocaleString()} tokens
+						</Label>
+					</LabelContainer>
+					<RangeInput
+						type="range"
+						min={MIN_VALID_TOKENS}
+						max={maxSliderValue}
+						step={100}
+						value={localValue}
+						onChange={handleSliderChange}
+						onMouseUp={handleSliderComplete}
+						onTouchEnd={handleSliderComplete}
+						$value={localValue}
+						$min={MIN_VALID_TOKENS}
+						$max={maxSliderValue}
+					/>
+
+					<Description>Higher budgets may allow you to achieve more comprehensive and nuanced reasoning</Description>
+				</>
+			)}
+		</Container>
+	)
+}
+
+export default memo(ThinkingBudgetSlider)
@@ -2,14 +2,7 @@ import React, { createContext, useCallback, useContext, useEffect, useState } fr
 import { useEvent } from "react-use"
 import { DEFAULT_AUTO_APPROVAL_SETTINGS } from "../../../src/shared/AutoApprovalSettings"
 import { ExtensionMessage, ExtensionState, DEFAULT_PLATFORM } from "../../../src/shared/ExtensionMessage"
-import {
-	ApiConfiguration,
-	ModelInfo,
-	openRouterDefaultModelId,
-	openRouterDefaultModelInfo,
-	requestyDefaultModelId,
-	requestyDefaultModelInfo,
-} from "../../../src/shared/api"
+import { ApiConfiguration, ModelInfo, openRouterDefaultModelId, openRouterDefaultModelInfo } from "../../../src/shared/api"
 import { findLastIndex } from "../../../src/shared/array"
 import { McpMarketplaceCatalog, McpServer } from "../../../src/shared/mcp"
 import { convertTextMateToHljs } from "../utils/textMateToHljs"
@@ -23,7 +16,6 @@ interface ExtensionStateContextType extends ExtensionState {
 	showWelcome: boolean
 	theme: any
 	openRouterModels: Record<string, ModelInfo>
-	requestyModels: Record<string, ModelInfo>
 	openAiModels: string[]
 	mcpServers: McpServer[]
 	mcpMarketplaceCatalog: McpMarketplaceCatalog
@@ -59,9 +51,6 @@ export const ExtensionStateContextProvider: React.FC<{
 	const [openRouterModels, setOpenRouterModels] = useState<Record<string, ModelInfo>>({
 		[openRouterDefaultModelId]: openRouterDefaultModelInfo,
 	})
-	const [requestyModels, setRequestyModels] = useState<Record<string, ModelInfo>>({
-		[requestyDefaultModelId]: requestyDefaultModelInfo,
-	})

 	const [openAiModels, setOpenAiModels] = useState<string[]>([])
 	const [mcpServers, setMcpServers] = useState<McpServer[]>([])
@@ -76,7 +65,6 @@ export const ExtensionStateContextProvider: React.FC<{
 					? [
 							config.apiKey,
 							config.openRouterApiKey,
-							config.requestyApiKey,
 							config.awsRegion,
 							config.vertexProjectId,
 							config.openAiApiKey,
@@ -91,6 +79,7 @@ export const ExtensionStateContextProvider: React.FC<{
 							config.qwenApiKey,
 							config.mistralApiKey,
 							config.vsCodeLmModelSelector,
+							config.asksageApiKey,
 							config.xaiApiKey,
 						].some((key) => key !== undefined)
 					: false
@@ -122,14 +111,6 @@ export const ExtensionStateContextProvider: React.FC<{
 				})
 				break
 			}
-			case "requestyModels": {
-				const updatedModels = message.requestyModels ?? {}
-				setRequestyModels({
-					[requestyDefaultModelId]: requestyDefaultModelInfo, // in case the extension sent a model list without the default model
-					...updatedModels,
-				})
-				break
-			}
 			case "openRouterModels": {
 				const updatedModels = message.openRouterModels ?? {}
 				setOpenRouterModels({
@@ -168,7 +149,6 @@ export const ExtensionStateContextProvider: React.FC<{
 		showWelcome,
 		theme,
 		openRouterModels,
-		requestyModels,
 		openAiModels,
 		mcpServers,
 		mcpMarketplaceCatalog,
@@ -5,13 +5,8 @@ import "./index.css"
 import App from "./App.tsx"
 import "../../node_modules/@vscode/codicons/dist/codicon.css"

-const apiKey = "phc_5WnLHpYyC30Bsb7VSJ6DzcPXZ34JSF08DJLyM7svZ15"
-const apiHost = "https://us.i.posthog.com"
-
 createRoot(document.getElementById("root")!).render(
 	<StrictMode>
-		<PostHogProvider apiKey={apiKey} options={{ api_host: apiHost }}>
-			<App />
-		</PostHogProvider>
+		<App />
 	</StrictMode>,
 )
@@ -180,7 +180,9 @@ export function getContextMenuOptions(
 		const seen = new Set()
 		const deduped = allItems.filter((item) => {
 			const key = `${item.type}-${item.value}`
-			if (seen.has(key)) return false
+			if (seen.has(key)) {
+				return false
+			}
 			seen.add(key)
 			return true
 		})
@@ -195,18 +197,26 @@ export function shouldShowContextMenu(text: string, position: number): boolean {
 	const beforeCursor = text.slice(0, position)
 	const atIndex = beforeCursor.lastIndexOf("@")

-	if (atIndex === -1) return false
+	if (atIndex === -1) {
+		return false
+	}

 	const textAfterAt = beforeCursor.slice(atIndex + 1)

 	// Check if there's any whitespace after the '@'
-	if (/\s/.test(textAfterAt)) return false
+	if (/\s/.test(textAfterAt)) {
+		return false
+	}

 	// Don't show the menu if it's a URL
-	if (textAfterAt.toLowerCase().startsWith("http")) return false
+	if (textAfterAt.toLowerCase().startsWith("http")) {
+		return false
+	}

 	// Don't show the menu if it's a problems or terminal
-	if (textAfterAt.toLowerCase().startsWith("problems") || textAfterAt.toLowerCase().startsWith("terminal")) return false
+	if (textAfterAt.toLowerCase().startsWith("problems") || textAfterAt.toLowerCase().startsWith("terminal")) {
+		return false
+	}

 	// NOTE: it's okay that menu shows when there's trailing punctuation since user could be inputting a path with marks

@@ -50,8 +50,10 @@ export const useShortcut = (shortcut: string, callback: any, options = { disable

 				if (Object.keys(modifierMap).includes(keyArray[0])) {
 					const finalKey = keyArray.pop()
+					if (!finalKey) return

-					if (keyArray.every((k) => modifierMap[k]) && finalKey === event.key) {
+					if (keyArray.every((k) => modifierMap[k]) && finalKey.toLowerCase() === event.key.toLowerCase()) {
+						event.preventDefault()
 						return callbackRef.current(event)
 					}
 				} else {
@@ -35,7 +35,9 @@ export function findMatchingResourceOrTemplate(
 ): McpResource | McpResourceTemplate | undefined {
 	// First try to find an exact resource match
 	const exactMatch = resources.find((resource) => resource.uri === uri)
-	if (exactMatch) return exactMatch
+	if (exactMatch) {
+		return exactMatch
+	}

 	// If no exact match, try to find a matching template
 	return findMatchingTemplate(uri, templates)
@@ -36,7 +36,5 @@ export function useDebounceEffect(effect: VoidFn, delay: number, deps: any[]) {

 		// We want to re‐schedule if any item in `deps` changed,
 		// or if `delay` changed.
-
-		// eslint-disable-next-line react-hooks/exhaustive-deps
 	}, [delay, ...deps])
 }
@@ -1,4 +1,4 @@
-import { ApiConfiguration, openRouterDefaultModelId, requestyDefaultModelId } from "../../../src/shared/api"
+import { ApiConfiguration, openRouterDefaultModelId } from "../../../src/shared/api"
 import { ModelInfo } from "../../../src/shared/api"
 export function validateApiConfiguration(apiConfiguration?: ApiConfiguration): string | undefined {
 	if (apiConfiguration) {
@@ -83,6 +83,11 @@ export function validateApiConfiguration(apiConfiguration?: ApiConfiguration): s
 					return "You must provide a valid model selector."
 				}
 				break
+			case "asksage":
+				if (!apiConfiguration.asksageApiKey) {
+					return "You must provide a valid API key or choose a different provider."
+				}
+				break
 		}
 	}
 	return undefined
@@ -91,26 +96,15 @@ export function validateApiConfiguration(apiConfiguration?: ApiConfiguration): s
 export function validateModelId(
 	apiConfiguration?: ApiConfiguration,
 	openRouterModels?: Record<string, ModelInfo>,
-	requestyModels?: Record<string, ModelInfo>,
 ): string | undefined {
 	if (apiConfiguration) {
 		switch (apiConfiguration.apiProvider) {
 			case "openrouter":
-				const openRouterModelId = apiConfiguration.openRouterModelId || openRouterDefaultModelId // in case the user hasn't changed the model id, it will be undefined by default
-				if (!openRouterModelId) {
+				const modelId = apiConfiguration.openRouterModelId || openRouterDefaultModelId // in case the user hasn't changed the model id, it will be undefined by default
+				if (!modelId) {
 					return "You must provide a model ID."
 				}
-				if (openRouterModels && !Object.keys(openRouterModels).includes(openRouterModelId)) {
-					// even if the model list endpoint failed, extensionstatecontext will always have the default model info
-					return "The model ID you provided is not available. Please choose a different model."
-				}
-				break
-			case "requesty":
-				const requestyModelId = apiConfiguration.requestyModelId || requestyDefaultModelId // in case the user hasn't changed the model id, it will be undefined by default
-				if (!requestyModelId) {
-					return "You must provide a model ID."
-				}
-				if (requestyModels && !Object.keys(requestyModels).includes(requestyModelId)) {
+				if (openRouterModels && !Object.keys(openRouterModels).includes(modelId)) {
 					// even if the model list endpoint failed, extensionstatecontext will always have the default model info
 					return "The model ID you provided is not available. Please choose a different model."
 				}
@@ -35,7 +35,9 @@ export function hexToRGB(hexColor: string): { r: number; g: number; b: number }

 export function colorToHex(colorVar: string): string {
 	const value = getComputedStyle(document.documentElement).getPropertyValue(colorVar).trim()
-	if (value.startsWith("#")) return value.slice(0, 7)
+	if (value.startsWith("#")) {
+		return value.slice(0, 7)
+	}

 	const rgbValues = value.match(/\d+/g)?.slice(0, 3).map(Number) || []
 	return `#${rgbValues.map((x) => x.toString(16).padStart(2, "0")).join("")}`
Author	SHA1	Message	Date
celestial-vault	5b537f8044	remove claude tool use backwards compatability	2025-03-04 21:49:48 -08:00
loupzeurandSaoud Rizwan	ccba2ed8f8	feat(ollama): use official ollama library (#1859 ) Co-authored-by: Saoud Rizwan <7799382+saoudrizwan@users.noreply.github.com>	2025-03-04 14:46:26 -08:00
Daniel Steigman	0fc2f474a4	Merge pull request #2075 from cline/improve-openrouter-error-message Openrouter: Typing Error Message	2025-03-03 22:11:44 -08:00
Yuto	0ce66efc19	Fixed problem with "win+shift+a" shortcut not working in Windows (#2049 ) * Fixed problem with “win+shift+a” shortcut not working in Windows * fix package-lock.json * Revert "fix package-lock.json" This reverts commit `1ffdfef4f3`.	2025-03-03 20:06:42 -08:00
Dennis Bartlett	e231cfd3e5	Feature/eng 238 (#2074 ) * Update linting to include webview. Fix outstanding lint issues. * Fix Format * Add changeset	2025-03-03 20:01:53 -08:00
celestial-vault	10f8637fb8	changeset	2025-03-03 17:47:10 -08:00
celestial-vault	bf985064b9	improve openrouter error types and pass extra error data if present	2025-03-03 17:37:59 -08:00
Saoud Rizwan	922183793f	Update lock	2025-03-02 17:23:45 -08:00
Saoud Rizwan	2c02547088	Copy	2025-03-02 17:22:20 -08:00
Saoud Rizwan	df65f5f5dd	Prepare for release	2025-03-02 17:19:56 -08:00
Saoud Rizwan	542ce36a80	Fix issue where thinking param was being sent with non-claude 3.7 models	2025-03-02 17:12:51 -08:00
Saoud Rizwan	6baac79138	Fix bedrock prompt caching	2025-03-02 17:11:11 -08:00
Saoud Rizwan	6a2172e156	Add gemini models to vertex	2025-03-01 18:03:30 -08:00
PaoandSaoud Rizwan	3ca529dc62	feat: Vertex Gemini Flash 2.0 support (#1853 ) * feat: Vertex Gemini Flash 2.0 support * fix: Add changeset for PR * Moved gemini model * Resolve conflicts * Prettier fix --------- Co-authored-by: Saoud Rizwan <7799382+saoudrizwan@users.noreply.github.com>	2025-03-01 17:55:43 -08:00
Saoud Rizwan	7728c0304c	Add o3-mini support to openai compatible	2025-03-01 17:35:04 -08:00
Saoud Rizwan	f837ecefeb	Remove border color	2025-03-01 16:51:28 -08:00
Saoud Rizwan	2d5dcf6a14	Add dcbartlett to codeowners	2025-03-01 16:48:06 -08:00
watany ellipsis-dev[bot]Saoud Rizwan	fcf6734949	feat(bedrock): Add APAC Support for Cross-Region Inference Profiles (#1913 ) * adding apac * Remove unreachable 'break' after return in case 'ap-'. Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com> * changeset * adding testing * revert * revart default model * revert package-lock * Remove unnecessary test --------- Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com> Co-authored-by: Saoud Rizwan <7799382+saoudrizwan@users.noreply.github.com>	2025-03-01 16:39:10 -08:00
Saoud Rizwan	d44fd44a81	Copy	2025-03-01 16:04:59 -08:00
Steven White Dennis Bartlett Saoud Rizwan	d103f8ee62	feat: add asksage support (#2011 ) * feat: add asksage support * chore: create changeset for asksage support * Fix Typo Fix Typo * chore: fix lint * Fixes * Validate asksage API key --------- Co-authored-by: Dennis Bartlett <bartlett.dc.1@gmail.com> Co-authored-by: Saoud Rizwan <7799382+saoudrizwan@users.noreply.github.com>	2025-03-01 15:20:58 -08:00
EvanandSaoud Rizwan	3d34de4908	Everyone Gets a Timeout (#2018 ) * re-open clean PR; formatting changes * remove unnecessary comments * Revert system prompt changes * Remove vscode message after timeout * Create tidy-kings-ring.md --------- Co-authored-by: Saoud Rizwan <7799382+saoudrizwan@users.noreply.github.com>	2025-03-01 14:46:29 -08:00
Roman Davydov	dcd275480f	Update recommendation from Claude 3.5 to Claude 3.7 (#2036 ) * Update Claude 3.5 -> Claude 3.7 in error message/recommendation * Add changeset	2025-03-01 14:22:18 -08:00
Dave Lasley	aea06ce659	Add retries to bedrock createMessage (#2039 ) * Add retries to bedrock createMessage Add retries to Bedrock's createMessage as indicated by [this](https://github.com/cline/cline/pull/1605#issuecomment-2665904349) Similar to [this comment](https://github.com/cline/cline/pull/1605#issuecomment-2666290702) I am also experiencing a 429 every once in a while. Measuring a before/after, it's about a 90% improvement when there's no prompt caching. I'm going to add some logging and see if I can track it down further but honestly this is probably worth it as is. * Add changeset	2025-03-01 14:20:50 -08:00
Leonid BugaevandSaoud Rizwan	65f582bb0d	Feat: Add support for AWS Bedrock Anthropic prompt caching (#2034 ) * Add Bedrock prompt caching support (optional) This feature protected under checkbox because it is not yet rolled out to everyone, and if you will try to send cache headers, and its not enabled for you, you will get error * Add changeset * Update supported models * Fix copy --------- Co-authored-by: Saoud Rizwan <7799382+saoudrizwan@users.noreply.github.com>	2025-03-01 14:09:10 -08:00
Saoud Rizwan	f8796ce69a	Fix changelog	2025-03-01 04:32:32 -08:00
Saoud Rizwan	4d29e1f4bd	Generate checkpoints asynchronously to prevent blocking UI	2025-03-01 04:19:37 -08:00
Saoud Rizwan	499c7dfb9f	Revert to legacy checkpoint tracking	2025-03-01 04:03:32 -08:00
Saoud Rizwan	57b36e16b8	Fix pricing calculation	2025-03-01 03:11:45 -08:00
Saoud Rizwan	c891d24ef6	Update budget description	2025-03-01 01:43:35 -08:00
Saoud Rizwan	3f5b6f6e33	Fix changelog	2025-03-01 01:40:42 -08:00
Saoud Rizwan	dda8a6c829	Add thinking budget option to Bedrock and Vertex	2025-03-01 01:39:14 -08:00
Saoud Rizwan	0959438db8	Add thinking budget slider to OpenRouter	2025-03-01 01:31:16 -08:00
Saoud Rizwan	a838cf9a61	Fix changelog	2025-03-01 00:57:07 -08:00
Saoud Rizwan	4e1dab8d4b	Fix changelog	2025-03-01 00:53:05 -08:00
Saoud Rizwan	93b6c937a5	Improve thinking budget slider animation; revert budget when switching plan/act modes	2025-03-01 00:49:34 -08:00
Saoud Rizwan	f555b83c7c	Revert "Sonnet: Make Separate Thinking Model (#2031 )" This reverts commit `29c01eb7d3`.	2025-02-28 23:04:20 -08:00
Saoud Rizwan	21d4ec13e1	Fix CHANGELOG	2025-02-28 22:52:03 -08:00
Saoud Rizwan	38e51bcf73	Update changelog	2025-02-28 22:51:08 -08:00
Saoud Rizwan	f40ff25d1e	Update package-locks	2025-02-28 22:45:13 -08:00
Evan	29c01eb7d3	Sonnet: Make Separate Thinking Model (#2031 ) * make thinking model separate and remove UI slider * use constant for min budget tokens * changeset	2025-02-28 22:36:02 -08:00
Saoud Rizwan	682fce458e	Prepare for release	2025-02-28 22:31:44 -08:00
Saoud Rizwan	34e87af5f1	Remove webview telemetry	2025-02-28 22:24:34 -08:00
Saoud Rizwan	eaa76512fc	Remove requesty polling; fix deepseek cost calculation changes; fix preferred language parsing	2025-02-28 22:23:24 -08:00
Evan	df7b458229	Sonnet, Give Me a Reason (#1961 ) * wip * added slider for setting reasoning budget tokens * refactor out generic slider component; improve styling; add debounce * added setting validation * styling and adding reasoning level * changeset * make change to trigger test rerun * revert useless comment	2025-02-28 18:16:56 -08:00
pashpashpash	560cefecca	Only requesting requesty models if requesty api key is set (#2028 ) * changeset * changeset	2025-02-28 17:59:21 -08:00
Dennis Bartlett	12dbf1556b	Revert Policy changes and TOS addition (#2027 )	2025-02-28 17:49:22 -08:00
pashpashpash	b4149b5a18	vertex prompt caching (#2026 ) * vertex prompt caching * duplicate declare * changeset	2025-02-28 17:30:25 -08:00
akfoster	cee959eeec	Merge pull request #2019 from cline/dev Merge dev into main	2025-02-28 17:08:06 -08:00
pashpashpash	46121094d7	package-lock.json	2025-02-28 16:56:46 -08:00
Frostbourne	6fe1015310	Update ClineProvider.ts	2025-02-28 14:26:08 -08:00
Dennis Bartlett	bb902f9b6b	Fix conflict	2025-02-28 14:24:51 -08:00
Dennis Bartlett	dfeb57d3af	Fix Lint	2025-02-28 14:06:00 -08:00
Dennis Bartlett	9232d752b2	Merge remote-tracking branch 'upstream/main' into dev	2025-02-28 14:05:27 -08:00
Saoud Rizwan	3525592c61	Prepare for release	2025-02-27 18:51:15 -08:00
Saoud Rizwan	1913b69ca5	Move GPT-4.5	2025-02-27 18:49:19 -08:00
Saoud Rizwan	aa714f28e1	Add GPT-4.5	2025-02-27 18:48:00 -08:00
Saoud Rizwan	eef04542d2	Revert package-lock	2025-02-27 18:40:37 -08:00