chore(tsconfig): normalize path aliases and exclude stories

- Remove deprecated `baseUrl` field and use explicit `./src`-based paths
- Align alias paths in root and webview tsconfigs for consistency
- Exclude Storybook stories and decorator from webview app build

.changeset/chubby-eels-look.md

@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Fix: do not retry request automatically on auth failure.

.changeset/famous-maps-scream.md

@@ -1,5 +0,0 @@
----
-"claude-dev": patch
----
-
-Added change to hide the context window usage message from env details when using next gen models and before the usage has reached an elevated state

.changeset/fifty-readers-cross.md

@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Fix task timeline display height.

.changeset/fix-contributing-proto-step.md

@@ -1,6 +0,0 @@
----
-"claude-dev": patch
----
-
-Docs: Add missing proto generation step in CONTRIBUTING.md and new `npm run dev` script for easier terminal workflow (fixes #7335)
-

.changeset/free-ducks-camp.md

@@ -1,5 +0,0 @@
----
-"claude-dev": patch
----
-
-Created model-family breakouts for deep-planning prompting, and laid groundwork for similar changes for other slash commands.

.changeset/free-nights-notice.md

@@ -1,5 +0,0 @@
----
-"claude-dev": patch
----
-
-Use HTTP proxies in more places

.changeset/fruity-wolves-fix.md

@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Ensure tool arguments are streamed during file operations when native tool calling is enabled.

.changeset/lemon-mammals-live.md

@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Added thinking level setting for Gemini 3.0 Pro

.changeset/light-buckets-wait.md

@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Feat: add thought signature support for Gemini SDK

.changeset/mighty-lizards-drop.md

@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Feat: Enable native tool calling for Baseten and Kimi K2 models

.changeset/shaky-places-allow.md

@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Replaces generic robot icon with Cline logo across VS Code UI

.changeset/shiny-flies-run.md

@@ -0,0 +1,5 @@
+---
+"claude-dev": minor
+---
+
+Add Kimi K2 Thinking to Baseten Provider

.changeset/solid-paths-battle.md

@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Skip MCP tool with invalid name (e.g. name too long) when native tool calling is enabled.

.changeset/swift-seas-work.md

@@ -1,5 +0,0 @@
----
-"claude-dev": patch
----
-
-Adjusted prompting around focus chain, particularly for next-get/native tool calling models.

.changeset/twenty-parents-taste.md

@@ -0,0 +1,5 @@
+---
+"claude-dev": patch
+---
+
+Fix Anthropic provider missing signature param when thinking is enabled.

.vscode/launch.json

@@ -165,6 +165,27 @@
 			},
 			"console": "integratedTerminal",
 			"internalConsoleOptions": "openOnSessionStart"
+		},
+		{
+			"name": "Open Storybook",
+			"type": "node",
+			"request": "launch",
+			"runtimeExecutable": "npm",
+			"runtimeArgs": [
+				"run",
+				"storybook"
+			],
+			"cwd": "${workspaceFolder}/webview-ui",
+			"console": "integratedTerminal",
+			"internalConsoleOptions": "neverOpen",
+			"serverReadyAction": {
+				"pattern": "Local:.*http://localhost:([0-9]+)",
+				"uriFormat": "http://localhost:%s",
+				"action": "openExternally"
+			},
+			"env": {
+				"IS_DEV": "true"
+			}
 		}
 	]
 }

.vscode/tasks.json

@@ -263,6 +263,26 @@
 				"watch"
 			],
 			"command": "rm -rf ${workspaceFolder}/dist/tmp/user && mkdir -p ${workspaceFolder}/dist/tmp/user"
+		},
+		{
+			"type": "npm",
+			"script": "storybook",
+			"group": "build",
+			"problemMatcher": [],
+			"isBackground": false,
+			"label": "npm: storybook",
+			"dependsOn": [
+				"npm: protos",
+				"npm: build:webview"
+			],
+			"presentation": {
+				"reveal": "always"
+			},
+			"options": {
+				"env": {
+					"IS_DEV": "true"
+				}
+			}
 		}
 	],
 	"inputs": [

.vscodeignore

@@ -20,6 +20,9 @@ eslint-rules/**
 .husky/**
 .env
 
+# cli
+cli/**
+
 # Custom
 **/demo.gif
 .nvmrc

CHANGELOG.md

@@ -1,5 +1,51 @@
 # Changelog
 
+## [3.38.1]
+
+### Fixed
+- Fixed handling of 'signature' field in sanitizeAnthropicContentBlock to properly preserve it when thinking is enabled, as required by Anthropic's API.
+
+## [3.38.0]
+
+### Added
+- Gemini 3 Pro Preview model
+- AquaVoice Avalon model for voice-to-text dictation
+
+### Fixed
+- Automatic context truncation when AWS Bedrock token usage rate limits are exceeded
+- Removed new_task tool from system prompts, updated slash command prompts, and added helper function for native tool calling validation
+
+
+## [3.37.1]
+
+- Comprehensive changes to better support GPT 5.1 - System prompt, tools, deep-planning, focus chain, etc.
+- Add AGENTS.md support
+- feat(models): Add free minimax/mimax-m2 model to the model picker
+
+## [3.37.0]
+
+### Added
+
+- GPT-5.1 with model-specific prompting: tailored system prompts, tool usage, focus chain, and deep-planning optimizations
+- Nous Research provider with Hermes 4 model family and custom system prompts
+- Switched to Aqua Voice's Avalon model in speech to text transcription
+- Added Linux support for speech to text
+- Model-family breakouts for deep-planning prompting, laying groundwork for enhanced slash commands
+- Expanded HTTP proxy support throughout the codebase
+- Improved focus chain prompting for frontier models (Anthropic, OpenAI, Gemini, xAI)
+
+### Fixed
+
+- Duplicate tool results prevention through existence checking
+- XML entity escaping in model content processor
+- Commit message generation in command palette
+- OpenAI Compatible provider temperature parameter type conversion
+
+## Documentation
+
+- Added missing proto generation step in CONTRIBUTING.md
+- New `npm run dev` script for streamlined terminal workflow (fixes #7335)
+
 ## [3.36.1]
 
 - fix: remove native tool calling support from Gemini and XAI provider due to invalid tool names issues

assets/icons/cline-bot.svg

@@ -0,0 +1,30 @@
+<?xml version="1.0" standalone="no"?>
+<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN" "http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd" >
+<svg xmlns="http://www.w3.org/2000/svg">
+<metadata>
+<json>
+<![CDATA[
+{
+	"fontFamily": "cline-bot",
+	"majorVersion": 1,
+	"minorVersion": 0,
+	"fontURL": "https://cline.bot",
+	"designerURL": "https://cline.bot",
+	"licenseURL": "https://cline.bot",
+	"version": "Version 1.0",
+	"fontId": "cline-bot",
+	"psName": "cline-bot",
+	"subFamily": "Regular",
+	"fullName": "cline-bot",
+	"description": "Font generated by IcoMoon."
+}
+]]>
+</json>
+</metadata>
+<defs>
+<font id="cline-bot" horiz-adv-x="1024">
+<font-face units-per-em="1024" ascent="960" descent="-64" />
+<missing-glyph horiz-adv-x="1024" />
+<glyph unicode="&#x20;" horiz-adv-x="512" d="" />
+<glyph unicode="&#xe900;" glyph-name="cline" data-tags="cline" horiz-adv-x="977" d="M964.553 383.11l-60.285 121.406v69.495c0 115.545-92.939 209.321-207.647 209.321h-102.986c7.536 15.071 11.722 32.654 11.722 51.074 0 64.471-51.912 116.383-115.545 116.383s-115.545-51.912-115.545-116.383 4.186-35.166 11.722-51.074h-102.986c-114.708 0-207.647-93.776-207.647-209.321v-69.495l-61.959-121.406c-5.861-11.722-5.861-26.793 0-38.515l61.959-119.732v-69.495c0-115.545 92.939-209.321 207.647-209.321h415.294c114.708 0 207.647 93.776 207.647 209.321v69.495l60.285 119.732c5.861 11.722 5.861 25.956 0 38.515v0zM426.178 284.311c0-52.749-42.702-95.451-94.613-95.451s-94.613 42.702-94.613 95.451v169.132c0 52.749 42.702 95.451 94.613 95.451s94.613-42.702 94.613-95.451v-169.132zM731.787 284.311c0-52.749-42.702-95.451-94.613-95.451s-94.613 42.702-94.613 95.451v169.132c0 52.749 42.702 95.451 94.613 95.451s94.613-42.702 94.613-95.451v-169.132z" />
+</font></defs></svg>

cli/pkg/cli/auth/byo_quick_setup.go

@@ -4,7 +4,6 @@ import (
 	"context"
 	"fmt"
 	"strings"
-	"time"
 
 	"github.com/cline/cli/pkg/cli/global"
 	"github.com/cline/cli/pkg/cli/task"
@@ -76,10 +75,11 @@ func QuickSetupFromFlags(ctx context.Context, provider, apiKey, modelID, baseURL
 		}
 	}
 
-	// WORKAROUND: Wait for debounced state persistence to complete
-	// Fixes `cline auth` issue when ran in docker environments
-	// TODO: implement better solution w/ changes in StateManager
-	time.Sleep(600 * time.Millisecond)
+	// Flush pending state changes to disk immediately
+	// This ensures all configuration changes are persisted before the instance terminates
+	if _, err := manager.GetClient().State.FlushPendingState(ctx, &cline.EmptyRequest{}); err != nil {
+		return fmt.Errorf("failed to flush pending state: %w", err)
+	}
 
 	// Success message
 	fmt.Printf("\n✓ Successfully configured %s provider\n", GetProviderDisplayName(providerEnum))
@@ -176,6 +176,7 @@ func validateQuickSetupProvider(providerID string) (cline.ApiProvider, error) {
 		cline.ApiProvider_XAI:           true,
 		cline.ApiProvider_CEREBRAS:      true,
 		cline.ApiProvider_OLLAMA:        true,
+		cline.ApiProvider_NOUSRESEARCH:  true,
 	}
 
 	if !supportedProviders[provider] {

cli/pkg/cli/auth/providers_byo.go

@@ -26,6 +26,7 @@ func GetBYOProviderList() []BYOProviderOption {
 		{Name: "Google Gemini", Provider: cline.ApiProvider_GEMINI},
 		{Name: "Ollama", Provider: cline.ApiProvider_OLLAMA},
 		{Name: "Cerebras", Provider: cline.ApiProvider_CEREBRAS},
+		{Name: "NousResearch", Provider: cline.ApiProvider_NOUSRESEARCH},
 		{Name: "Oracle Code Assist", Provider: cline.ApiProvider_OCA},
 	}
 }
@@ -100,6 +101,8 @@ func GetBYOProviderPlaceholder(provider cline.ApiProvider) string {
 		return "e.g., qwen3-coder:30b"
 	case cline.ApiProvider_CEREBRAS:
 		return "e.g., gpt-oss-120b"
+	case cline.ApiProvider_NOUSRESEARCH:
+		return "e.g., Hermes-4-405B"
 	case cline.ApiProvider_OCA:
 		return "e.g., oca/llama4"
 	default:

cli/pkg/cli/auth/providers_list.go

@@ -111,6 +111,7 @@ func (r *ProviderListResult) GetAllReadyProviders() []*ProviderDisplay {
 		cline.ApiProvider_GEMINI,
 		cline.ApiProvider_OLLAMA,
 		cline.ApiProvider_CEREBRAS,
+		cline.ApiProvider_NOUSRESEARCH,
 		cline.ApiProvider_OCA,
 		cline.ApiProvider_HICAP,
 	}
@@ -241,6 +242,8 @@ func mapProviderStringToEnum(providerStr string) (cline.ApiProvider, bool) {
 		return cline.ApiProvider_OCA, true
 	case "hicap":
 		return cline.ApiProvider_HICAP, true
+	case "nousResearch":
+		return cline.ApiProvider_NOUSRESEARCH, true
 	default:
 		return cline.ApiProvider_ANTHROPIC, false // Return 0 value with false
 	}
@@ -274,6 +277,8 @@ func GetProviderIDForEnum(provider cline.ApiProvider) string {
 		return "oca"
 	case cline.ApiProvider_HICAP:
 		return "hicap"
+	case cline.ApiProvider_NOUSRESEARCH:
+		return "nousResearch"
 	default:
 		return ""
 	}
@@ -353,6 +358,8 @@ func GetProviderDisplayName(provider cline.ApiProvider) string {
 		return "Oracle Code Assist"
 	case cline.ApiProvider_HICAP:
 		return "Hicap"
+	case cline.ApiProvider_NOUSRESEARCH:
+		return "NousResearch"
 	default:
 		return "Unknown"
 	}
@@ -475,6 +482,7 @@ func DetectAllConfiguredProviders(ctx context.Context, manager *task.Manager) ([
 		{cline.ApiProvider_OLLAMA, "ollamaBaseUrl"}, // Ollama uses baseUrl instead of API key
 		{cline.ApiProvider_CEREBRAS, "cerebrasApiKey"},
 		{cline.ApiProvider_HICAP, "hicapApiKey"},
+		{cline.ApiProvider_NOUSRESEARCH, "nousResearchApiKey"},
 	}
 
 	for _, providerCheck := range providersToCheck {

cli/pkg/cli/auth/update_api_configurations.go

@@ -163,6 +163,15 @@ func GetProviderFields(provider cline.ApiProvider) (ProviderFields, error) {
 			ActModeProviderSpecificModelIDField:  "actModeHicapModelId",
 		}, nil
 
+	case cline.ApiProvider_NOUSRESEARCH:
+		return ProviderFields{
+			APIKeyField:                          "nousResearchApiKey",
+			PlanModeModelIDField:                 "planModeApiModelId",
+			ActModeModelIDField:                  "actModeApiModelId",
+			PlanModeProviderSpecificModelIDField: "planModeNousResearchModelId",
+			ActModeProviderSpecificModelIDField:  "actModeNousResearchModelId",
+		}, nil
+
 	default:
 		return ProviderFields{}, fmt.Errorf("unsupported provider: %v", provider)
 	}
@@ -278,6 +287,8 @@ func setAPIKeyField(apiConfig *cline.ModelsApiConfiguration, fieldName string, v
 		apiConfig.OcaApiKey = value
 	case "hicapApiKey":
 		apiConfig.HicapApiKey = value
+	case "nousResearchApiKey":
+		apiConfig.NousResearchApiKey = value
 	}
 }
 
@@ -302,6 +313,9 @@ func setProviderSpecificModelID(apiConfig *cline.ModelsApiConfiguration, fieldNa
 	case "planModeHicapModelId":
 		apiConfig.PlanModeHicapModelId = value
 		apiConfig.ActModeHicapModelId = value
+	case "planModeNousResearchModelId":
+		apiConfig.PlanModeNousResearchModelId = value
+		apiConfig.ActModeNousResearchModelId = value
 	}
 }
 

cli/pkg/cli/auth/wizard_byo.go

@@ -517,7 +517,7 @@ func (pw *ProviderWizard) applyModelChange(provider cline.ApiProvider, modelID s
 		ModelInfo: modelInfo,
 	}
 
-	return UpdateProviderPartial(pw.ctx, pw.manager, provider, updates, false)
+	return UpdateProviderPartial(pw.ctx, pw.manager, provider, updates, true)
 }
 
 // SwitchToBYOProvider switches to a BYO provider that's already configured.

cli/pkg/cli/display/tool_result_parser.go

@@ -221,83 +221,7 @@ func (p *ToolResultParser) ParseCodeDefinitions(content string) string {
 
 // ParseWebFetch formats webFetch tool results with content preview
 func (p *ToolResultParser) ParseWebFetch(content, url string) string {
-	if content == "" {
-		return fmt.Sprintf("*Fetched content from %s (empty response)*", url)
-	}
-
-	lines := strings.Split(content, "\n")
-	
-	var result strings.Builder
-	
-	// Try to extract title
-	var title string
-	for _, line := range lines {
-		trimmed := strings.TrimSpace(line)
-		if strings.HasPrefix(trimmed, "#") && !strings.HasPrefix(trimmed, "##") {
-			title = strings.TrimSpace(strings.TrimPrefix(trimmed, "#"))
-			break
-		}
-	}
-
-	if title != "" {
-		result.WriteString(fmt.Sprintf("**Title:** %s\n\n", title))
-	}
-
-	// Show preview of content
-	result.WriteString("**Preview:**\n")
-	
-	charCount := 0
-	maxChars := 500
-	previewLines := []string{}
-	
-	for _, line := range lines {
-		// Skip markdown headers
-		if strings.HasPrefix(strings.TrimSpace(line), "#") {
-			continue
-		}
-		
-		trimmed := strings.TrimSpace(line)
-		if trimmed == "" {
-			continue
-		}
-		
-		if charCount+len(trimmed) > maxChars {
-			break
-		}
-		
-		previewLines = append(previewLines, trimmed)
-		charCount += len(trimmed)
-	}
-
-	result.WriteString(strings.Join(previewLines, " "))
-	result.WriteString("...\n\n")
-
-	// Extract sections
-	sections := []string{}
-	for _, line := range lines {
-		trimmed := strings.TrimSpace(line)
-		if strings.HasPrefix(trimmed, "##") {
-			section := strings.TrimSpace(strings.TrimPrefix(trimmed, "##"))
-			sections = append(sections, section)
-			if len(sections) >= 5 {
-				break
-			}
-		}
-	}
-
-	if len(sections) > 0 {
-		result.WriteString("**Sections Found:**\n")
-		for _, section := range sections {
-			result.WriteString(fmt.Sprintf("- %s\n", section))
-		}
-		result.WriteString("\n")
-	}
-
-	// Word count estimate
-	wordCount := len(strings.Fields(content))
-	result.WriteString(fmt.Sprintf("*[Full content: ~%s]*", p.formatWordCount(wordCount)))
-
-	return result.String()
+	return ""
 }
 
 // detectLanguage returns syntax highlighting language based on file extension

cli/pkg/generated/providers.go

@@ -145,6 +145,7 @@ const (
 	XAI = "xai"
 	CEREBRAS = "cerebras"
 	OCA = "oca"
+	NOUSRESEARCH = "nousResearch"
 )
 
 // AllProviders returns a slice of enabled provider IDs for the CLI build.
@@ -161,6 +162,7 @@ var AllProviders = []string{
 	"xai",
 	"cerebras",
 	"oca",
+	"nousResearch",
 }
 
 // ConfigField represents a configuration field requirement
@@ -318,6 +320,15 @@ var rawConfigFields = `	[
 	    "fieldType": "password",
 	    "placeholder": "Enter your API key"
 	  },
+	  {
+	    "name": "nousResearchApiKey",
+	    "type": "string",
+	    "comment": "",
+	    "category": "nousResearch",
+	    "required": true,
+	    "fieldType": "password",
+	    "placeholder": "Enter your API key"
+	  },
 	  {
 	    "name": "ulid",
 	    "type": "string",
@@ -435,6 +446,15 @@ var rawConfigFields = `	[
 	    "fieldType": "url",
 	    "placeholder": "https://api.example.com"
 	  },
+	  {
+	    "name": "minimaxApiLine",
+	    "type": "string",
+	    "comment": "",
+	    "category": "general",
+	    "required": false,
+	    "fieldType": "string",
+	    "placeholder": ""
+	  },
 	  {
 	    "name": "ocaMode",
 	    "type": "string",
@@ -775,6 +795,24 @@ var rawModelDefinitions = `	{
 	      "supportsImages": false,
 	      "supportsPromptCache": false,
 	      "description": "A compact 20B open-weight Mixture-of-Experts language model designed for strong reasoning and tool use, ideal for edge devices and local inference."
+	    },
+	    "qwen.qwen3-coder-30b-a3b-v1:0": {
+	      "maxTokens": 8192,
+	      "contextWindow": 262144,
+	      "inputPrice": 0,
+	      "outputPrice": 0,
+	      "supportsImages": false,
+	      "supportsPromptCache": false,
+	      "description": "Qwen3 Coder 30B MoE model with 3.3B activated parameters, optimized for code generation and analysis with 256K context window."
+	    },
+	    "qwen.qwen3-coder-480b-a35b-v1:0": {
+	      "maxTokens": 8192,
+	      "contextWindow": 262144,
+	      "inputPrice": 0,
+	      "outputPrice": 1,
+	      "supportsImages": false,
+	      "supportsPromptCache": false,
+	      "description": "Qwen3 Coder 480B flagship MoE model with 35B activated parameters, designed for complex coding tasks with advanced reasoning capabilities and 256K context window."
 	    }
 	  },
 	  "gemini": {
@@ -1263,6 +1301,26 @@ var rawModelDefinitions = `	{
 	      "supportsPromptCache": false,
 	      "description": "SOTA performance with ~1500 tokens/s"
 	    }
+	  },
+	  "nousResearch": {
+	    "Hermes-4-405B": {
+	      "maxTokens": 8192,
+	      "contextWindow": 128000,
+	      "inputPrice": 0,
+	      "outputPrice": 0,
+	      "supportsImages": false,
+	      "supportsPromptCache": false,
+	      "description": "This is the largest model in the Hermes 4 family, and it is the fullest expression of our design, focused on advanced reasoning and creative depth rather than optimizing inference speed or cost."
+	    },
+	    "Hermes-4-70B": {
+	      "maxTokens": 8192,
+	      "contextWindow": 128000,
+	      "inputPrice": 0,
+	      "outputPrice": 0,
+	      "supportsImages": false,
+	      "supportsPromptCache": false,
+	      "description": "This incarnation of Hermes 4 balances scale and size. It handles complex reasoning tasks, while staying fast and cost effective. A versatile choice for many use cases."
+	    }
 	  }
 	}`
 
@@ -1432,6 +1490,18 @@ func GetProviderDefinitions() (map[string]ProviderDefinition, error) {
 		HasDynamicModels: false,
 		SetupInstructions: `Configure Oca API credentials`,
 	}
+
+	// NousResearch
+	definitions["nousResearch"] = ProviderDefinition{
+		ID:              "nousResearch",
+		Name:            "NousResearch",
+		RequiredFields:  getFieldsByProvider("nousResearch", configFields, true),
+		OptionalFields:  getFieldsByProvider("nousResearch", configFields, false),
+		Models:          modelDefinitions["nousResearch"],
+		DefaultModelID:  "Hermes-4-405B",
+		HasDynamicModels: false,
+		SetupInstructions: `Configure NousResearch API credentials`,
+	}
 	
 	return definitions, nil
 }
@@ -1459,6 +1529,7 @@ func GetProviderDisplayName(providerID string) string {
 		"xai": "X AI (Grok)",
 		"cerebras": "Cerebras",
 		"oca": "Oca",
+		"nousResearch": "NousResearch",
 	}
 	
 	if name, exists := displayNames[providerID]; exists {

docs/docs.json

@@ -189,6 +189,7 @@
 									"provider-config/fireworks",
 									"provider-config/zai",
 									"provider-config/gcp-vertex-ai",
+									"provider-config/baseten",
 									{
 										"group": "AWS Bedrock",
 										"pages": [
@@ -215,8 +216,7 @@
 									"provider-config/vscode-language-model-api",
 									"provider-config/sap-aicore",
 									"provider-config/vercel-ai-gateway",
-									"provider-config/requesty",
-									"provider-config/baseten"
+									"provider-config/requesty"
 								]
 							}
 						]

docs/features/cline-rules.mdx

@@ -81,6 +81,20 @@ your-project/
 
 Cline's system prompt, on the other hand, is not user-editable ([here's where you can find it](https://github.com/cline/cline/blob/main/src/core/prompts/system.ts)). For a broader look at prompt engineering best practices, check out [this resource](https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/overview).
 
+### AGENTS.md Standard Support
+
+Cline also supports the [AGENTS.md](https://agents.md/) standard as a fallback 
+(in addition to Cline Rules) by automatically detecting `AGENTS.md` files in 
+your workspace root. This allows you to use the same rules file across different AI 
+coding tools.
+
+```
+your-project/
+├── AGENTS.md
+├── src/
+└── ...
+```
+
 ### Tips for Writing Effective Cline Rules
 
 -   Be Clear and Concise: Use simple language and avoid ambiguity.

docs/features/dictation.mdx

@@ -3,7 +3,7 @@ title: "Dictation"
 description: "Communicate with Cline using your voice for faster, more natural AI collaboration"
 ---
 
-Dictation transforms how you work with AI. Instead of typing out complex thoughts, you speak naturally and share your complete intent. This isn't just about speed - though voice is faster - it's about unlocking the kind of fluid collaboration that typing can't match.
+Dictation transforms how you work with AI. Instead of typing out complex thoughts, you speak naturally and share your complete intent. This isn't just about speed - though voice is faster - it's about enabling fluid collaboration that typing can't match.
 
 ## Why Voice Changes Everything
 
@@ -35,11 +35,14 @@ Dictation works with any AI model you've configured. The transcription happens t
 
 ## System Requirements
 
+<Note>
+Dictation is currently not available on Windows. Support for Windows is planned for a future release.
+</Note>
+
 Dictation uses FFmpeg to capture your voice across all platforms:
 
 - **macOS**: FFmpeg (via Homebrew: `brew install ffmpeg`)
 - **Linux**: FFmpeg (via apt: `sudo apt-get install ffmpeg`)
-- **Windows**: FFmpeg (via winget: `winget install Gyan.FFmpeg`)
 
 If you don't have FFmpeg installed, Cline will automatically detect this and prompt you to install it with a single click.
 

docs/features/hooks.mdx

@@ -76,7 +76,7 @@ echo "$input" | jq -r '.timestamp | type'
 
 This example script demonstrates the key mechanics of hook input/output: reading the JSON payload from stdin with `input=$(cat)`, and using `jq` to inspect the data structure and field types that your hook receives. This helps you understand what data is available before building more complex hook logic.
 
-#### Make it executable
+**Make it executable**
 
 ```bash
 chmod +x .clinerules/hooks/TaskStart
@@ -92,6 +92,32 @@ Start a task in Cline and verify your hook executes.
 Start with a simple hook that just logs information before building complex validation logic. This helps you understand the data structure and timing.
 </Tip>
 
+
+## What You Can Build
+
+Once you understand the basics, hooks open up creative possibilities:
+
+<CardGroup cols={2}>
+  <Card title="Intelligent Code Review" icon="code-branch">
+    Run linters or custom validators before files get saved. Block commits that don't pass checks. Track code quality metrics over time.
+  </Card>
+  
+  <Card title="Security Enforcement" icon="shield-halved">
+    Prevent operations that violate security policies. Detect when sensitive data might be exposed. Audit all file access for compliance.
+  </Card>
+  
+  <Card title="Development Analytics" icon="chart-line">
+    Measure how long different operations take. Identify patterns in how the AI works. Generate productivity reports from hook data.
+  </Card>
+  
+  <Card title="Integration Hub" icon="plug">
+    Connect to issue trackers when certain keywords appear. Update project management tools. Sync with external APIs at the right moments.
+  </Card>
+</CardGroup>
+
+The key is combining hooks with external tools. A hook can be the glue between Cline's workflow and the rest of your development ecosystem.
+
+
 ## Hook Types
 
 Cline provides multiple hook types that let you tap into different stages of the AI workflow. They're organized into categories based on their trigger points and use cases.
@@ -341,22 +367,6 @@ Context injection affects future decisions, not current ones. When a hook runs:
 This means PreToolUse hooks are for blocking bad actions, while PostToolUse hooks are for learning from completed ones.
 
 
-## What You Can Build
-
-Once you understand the basics, hooks open up creative possibilities:
-
-- **Intelligent Code Review**:
-Run linters or custom validators before files get saved. Block commits that don't pass checks. Track code quality metrics over time.
-
-- **Security Enforcement**:
-Prevent operations that violate security policies. Detect when sensitive data might be exposed. Audit all file access for compliance.
-
-- **Development Analytics**: Measure how long different operations take. Identify patterns in how the AI works. Generate productivity reports from hook data.
-
-- **Integration Hub**: Connect to issue trackers when certain keywords appear. Update project management tools. Sync with external APIs at the right moments.
-
-The key is combining hooks with external tools. A hook can be the glue between Cline's workflow and the rest of your development ecosystem.
-
 ## Troubleshooting
 
 ### Hook Not Running
@@ -371,12 +381,30 @@ The key is combining hooks with external tools. A hook can be the glue between C
 - Consider moving complex logic to a background process
 
 ### Context Not Affecting Behavior
-- Remember: context affects FUTURE decisions, not the current tool
-- The current AI behavior is based on the previous "API Request..." block
-- Your `contextModification` gets injected into the NEXT "API Request..." block
-- Use PreToolUse for validation (blocking) if you need immediate effect
-- Ensure context modifications are clear and actionable
-- Check that context isn't being truncated (50KB limit)
+Remember that context modifications affect future AI decisions, not the current operation. The AI's current behavior is based on the previous "API Request..." block, and your `contextModification` gets injected into the next "API Request..." block. This means if you need immediate effect, you should use PreToolUse hooks for validation and return `cancel: true` in your hook's JSON response to block Cline from continuing.
+
+When adding context, ensure your modifications are clear and actionable so the AI can understand and apply them effectively. Also check that your context isn't being truncated due to the 50KB limit, as this could prevent important information from reaching the AI.
+
+### Handling Strings with Quotes in JSON Payloads
+When your hook needs to include strings containing unescaped quote characters (`"`) in JSON output, use jq's `--arg` flag for proper escaping:
+
+```bash
+#!/usr/bin/env bash
+
+# When $output contains unescaped quote characters (")...
+output='{"foo":"bar"}'
+
+# Use the --arg flag for automatic string escaping
+jq -n --arg ctx "$output" '{cancel: false, contextModification: $ctx}'
+
+# This will result in:
+# {
+#     "cancel": false,
+#     "contextModification": "{\"foo\":\"bar\"}"
+# }
+```
+
+The `--arg` flag automatically escapes special characters, preventing JSON parsing errors when your context modification includes complex strings or nested JSON structures.
 
 <Warning>
 Hooks run with the same permissions as VS Code. They can access all workspace files and environment variables. Review hooks from untrusted sources before enabling them.

docs/mcp/connecting-to-a-remote-server.mdx

@@ -29,11 +29,33 @@ The "Remote Servers" tab allows you to connect to any MCP server that's accessib
 2. Fill in the required information:
     - **Server Name**: Provide a unique, descriptive name for the server
     - **Server URL**: Enter the complete URL endpoint of the MCP server (e.g., `https://example.com/mcp-sse`)
+    - **Transport Type**: Select the connection protocol (Streamable HTTP is recommended for modern servers)
 3. Click "Add Server" to initiate the connection
 4. Cline will attempt to connect to the server and display the connection status
 
 > **Note**: When connecting to a remote server, ensure you trust the source, as MCP servers can execute code in your environment.
 
+#### Transport Types
+
+Cline supports two transport protocols for remote MCP servers:
+
+-   **Streamable HTTP (Recommended)**: The modern MCP transport protocol with better performance, reliability, and full OAuth 2.1 authentication support. Use this for most remote servers.
+-   **SSE (Legacy)**: Server-Sent Events transport. Use this only if the server specifically requires SSE or doesn't support Streamable HTTP.
+
+#### OAuth Authentication
+
+Some MCP servers (like Vercel's MCP) require OAuth authentication to access your data securely. When connecting to an OAuth-enabled server:
+
+1. Add the server as usual with its URL
+2. If the server requires authentication, you'll see an error message asking to authenticate.
+3. Click the **"Authenticate"** button that appears
+4. Your browser will open to the server's authorization page
+5. Sign in and grant permission
+6. You'll be redirected back to Cline automatically
+7. The server will connect and show a green status dot
+
+Once authenticated, your credentials are securely stored and the server will reconnect automatically when you reload Cline. You won't need to authenticate again unless you delete the server or your credentials expire.
+
 ### Remote Server Discovery
 
 If you're looking for MCP servers to connect to, several third-party marketplaces provide directories of available servers with various capabilities.
@@ -90,9 +112,20 @@ Toggle the switch next to each server to enable or disable it:
 If a server fails to connect:
 
 1. An error message will be displayed with details about the failure
-2. Check that the server URL is correct and the server is running
-3. Use the "Restart Server" button to attempt reconnection
-4. If problems persist, you can delete the server and try adding it again
+2. **For OAuth errors**: Click the "Authenticate" button to complete the authorization flow
+3. Check that the server URL is correct and the server is running
+4. Try selecting a different transport type (Streamable HTTP vs SSE)
+5. Use the "Restart Server" button to attempt reconnection
+6. If problems persist, you can delete the server and try adding it again
+
+#### OAuth-Specific Issues
+
+If you're having trouble authenticating with an OAuth-enabled server:
+
+-   **"Authentication required" persists**: Make sure you completed the authorization flow in your browser and didn't cancel it
+-   **Browser doesn't open**: Check your system's default browser settings and ensure external URLs can be opened
+-   **Redirect errors**: Verify you're using the latest version of Cline - older versions may not support OAuth
+-   **Reset authentication**: Delete the server and re-add it to start fresh with a new OAuth flow
 
 ### Advanced Configuration
 
@@ -105,10 +138,11 @@ For advanced users, Cline stores MCP server configurations in a JSON file that c
 {
 	"mcpServers": {
 		"exampleServer": {
-			"url": "https://example.com/mcp-sse",
+			"url": "https://example.com/mcp-server",
+			"type": "streamableHttp",
 			"disabled": false,
 			"autoApprove": ["tool1", "tool2"],
-			"timeout": 30
+			"timeout": 60
 		}
 	}
 }
@@ -117,9 +151,10 @@ For advanced users, Cline stores MCP server configurations in a JSON file that c
 Key configuration options:
 
 -   **url**: The endpoint URL (for remote servers)
+-   **type**: Transport protocol - `"streamableHttp"` (recommended) or `"sse"` (legacy)
 -   **disabled**: Whether the server is currently enabled (true/false)
 -   **autoApprove**: List of tool names that don't require confirmation
--   **timeout**: Maximum time in seconds to wait for server responses
+-   **timeout**: Maximum time in seconds to wait for server responses (default: 60)
 
 For additional MCP settings, click the "Advanced MCP Settings" link to access VSCode settings.
 
@@ -130,3 +165,20 @@ Once connected, Cline can use the tools and resources provided by the MCP server
 1. A tool approval prompt will appear (unless auto-approved)
 2. Review the tool details and parameters before approving
 3. The tool will execute and return results to Cline
+
+### Example: Connecting to Vercel MCP
+
+[Vercel MCP](https://vercel.com/docs/mcp/vercel-mcp) is an OAuth-enabled server that provides tools for managing your Vercel projects and deployments:
+
+1. Click "Remote Servers" tab
+2. Enter:
+    - **Server Name**: `vercel`
+    - **Server URL**: `https://mcp.vercel.com`
+    - **Transport Type**: Streamable HTTP (pre-selected)
+3. Click "Add Server"
+4. You'll see "Authentication required" - click the **"Authenticate"** button
+5. Sign in to Vercel in your browser and authorize Cline
+6. Return to Cline - the server will automatically connect
+7. Vercel's tools (deploy, logs, projects) are now available to Cline!
+
+Your Vercel authentication persists across sessions, so you won't need to re-authenticate each time you use Cline.

docs/package-lock.json

@@ -5146,28 +5146,6 @@
 				"node": ">=6.0"
 			}
 		},
-		"node_modules/gray-matter/node_modules/argparse": {
-			"version": "1.0.10",
-			"resolved": "https://registry.npmjs.org/argparse/-/argparse-1.0.10.tgz",
-			"integrity": "sha512-o5Roy6tNG4SL/FOkCAN6RzjiakZS25RLYFrcMttJqbdd8BWrnA+fGz57iN5Pb06pvBGvl5gQ0B48dJlslXvoTg==",
-			"license": "MIT",
-			"dependencies": {
-				"sprintf-js": "~1.0.2"
-			}
-		},
-		"node_modules/gray-matter/node_modules/js-yaml": {
-			"version": "3.14.1",
-			"resolved": "https://registry.npmjs.org/js-yaml/-/js-yaml-3.14.1.tgz",
-			"integrity": "sha512-okMH7OXXJ7YrN9Ok3/SXrnu4iX9yOk+25nqX4imS2npuvTYDmo/QEZoqwZkYaIDk3jVvBOTOIEgEhaLOynBS9g==",
-			"license": "MIT",
-			"dependencies": {
-				"argparse": "^1.0.7",
-				"esprima": "^4.0.0"
-			},
-			"bin": {
-				"js-yaml": "bin/js-yaml.js"
-			}
-		},
 		"node_modules/has-bigints": {
 			"version": "1.1.0",
 			"resolved": "https://registry.npmjs.org/has-bigints/-/has-bigints-1.1.0.tgz",
@@ -6490,9 +6468,9 @@
 			"license": "MIT"
 		},
 		"node_modules/js-yaml": {
-			"version": "4.1.0",
-			"resolved": "https://registry.npmjs.org/js-yaml/-/js-yaml-4.1.0.tgz",
-			"integrity": "sha512-wpxZs9NoxZaJESJGIZTyDEaYpl0FKSA+FB9aJiyemKhMwkxQg63h4T1KJgUGHpTqPDNRcmmYLugrRjJlBtWvRA==",
+			"version": "4.1.1",
+			"resolved": "https://registry.npmjs.org/js-yaml/-/js-yaml-4.1.1.tgz",
+			"integrity": "sha512-qQKT4zQxXl8lLwBtHMWwaTcGfFOZviOJet3Oy/xmGk2gZH677CJM9EvtfdSkgWcATZhj/55JZ0rmy3myCT5lsA==",
 			"license": "MIT",
 			"dependencies": {
 				"argparse": "^2.0.1"
@@ -10235,12 +10213,6 @@
 				"url": "https://github.com/sponsors/wooorm"
 			}
 		},
-		"node_modules/sprintf-js": {
-			"version": "1.0.3",
-			"resolved": "https://registry.npmjs.org/sprintf-js/-/sprintf-js-1.0.3.tgz",
-			"integrity": "sha512-D9cPgkvLlV3t3IzL0D0YLvGA9Ahk4PcvVwUbN0dSGr1aP0Nrt4AEnTUbuGvquEC0mA64Gqt1fzirlRs5ibXx8g==",
-			"license": "BSD-3-Clause"
-		},
 		"node_modules/stack-utils": {
 			"version": "2.0.6",
 			"resolved": "https://registry.npmjs.org/stack-utils/-/stack-utils-2.0.6.tgz",
@@ -10609,9 +10581,9 @@
 			}
 		},
 		"node_modules/tar-fs": {
-			"version": "3.1.0",
-			"resolved": "https://registry.npmjs.org/tar-fs/-/tar-fs-3.1.0.tgz",
-			"integrity": "sha512-5Mty5y/sOF1YWj1J6GiBodjlDc05CUR8PKXrsnFAiSG0xA+GHeWLovaZPYUDXkH/1iKRf2+M5+OrRgzC7O9b7w==",
+			"version": "3.1.1",
+			"resolved": "https://registry.npmjs.org/tar-fs/-/tar-fs-3.1.1.tgz",
+			"integrity": "sha512-LZA0oaPOc2fVo82Txf3gw+AkEd38szODlptMYejQUhndHMLQ9M059uXR+AfS7DNo0NpINvSqDsvyaCrBVkptWg==",
 			"license": "MIT",
 			"dependencies": {
 				"pump": "^3.0.0",

docs/package.json

@@ -14,5 +14,9 @@
 	"description": "",
 	"dependencies": {
 		"mintlify": "^4.2.23"
+	},
+	"overrides": {
+		"tar-fs": "^3.1.1",
+		"js-yaml": "^4.1.1"
 	}
 }

docs/provider-config/baseten.mdx

@@ -3,7 +3,7 @@ title: "Baseten"
 description: "Learn how to configure and use Baseten's Model APIs with Cline. Access frontier open-source models with enterprise-grade performance, reliability, and competitive pricing."
 ---
 
-Baseten provides on-demand frontier model APIs designed for production applications, not just experimentation. Built on the Baseten Inference Stack, these APIs deliver enterprise-grade performance and reliability with optimized inference for leading open-source models from OpenAI, DeepSeek, Meta, Moonshot AI, and Alibaba Cloud.
+Baseten provides on-demand frontier model APIs designed for production applications, not just experimentation. Built on the Baseten Inference Stack, these APIs deliver optimized inference for leading open-source models from OpenAI, DeepSeek, Moonshot AI, and Alibaba Cloud.
 
 **Website:** [https://www.baseten.co/products/model-apis/](https://www.baseten.co/products/model-apis/)
 
@@ -14,13 +14,21 @@ Baseten provides on-demand frontier model APIs designed for production applicati
 3.  **Create a Key:** Generate a new API key. Give it a descriptive name (e.g., "Cline").
 4.  **Copy the Key:** Copy the API key immediately and store it securely.
 
+### Configuration in Cline
+
+1.  **Open Cline Settings:** Click the settings icon (⚙️) in the Cline panel.
+2.  **Select Provider:** Choose "Baseten" from the "API Provider" dropdown.
+3.  **Enter API Key:** Paste your Baseten API key into the "Baseten API Key" field.
+4.  **Select Model:** Choose your desired model from the "Model" dropdown.
+
+**IMPORTANT: For Kimi K2 Thinking:** To use the `moonshotai/Kimi-K2-Thinking` model, you must enable **Native Tool Call (Experimental)** in Cline settings. This setting allows Cline to call tools through their native tool processor and is required for this reasoning model to function properly.
+
 ### Supported Models
 
 Cline supports all current models under Baseten Model APIs, including:
 For the most updated pricing, please visit: https://www.baseten.co/products/model-apis/
-Note: Kimi K2 0711, Llama 4 Maverick, and Llama 4 Scout Model APIs have been deprecated at 5pm PT on October 8th.
-https://www.baseten.co/resources/changelog/model-api-deprecation-notice-kimi-k2-0711-scout-maverick/
 
+-   `moonshotai/Kimi-K2-Thinking` (Moonshot AI) - Enhanced reasoning capabilities with step-by-step thought processes (262K context) - \$0.60/\$2.50 per 1M tokens
 -   `zai-org/GLM-4.6` (Z AI) - Frontier open model with advanced agentic, reasoning and coding capabilities by Z AI (200k context) \$0.60/\$2.20 per 1M tokens
 -   `moonshotai/Kimi-K2-Instruct-0905` (Moonshot AI) - September update with enhanced capabilities (262K context) - \$0.60/\$2.50 per 1M tokens
 -   `openai/gpt-oss-120b` (OpenAI) - 120B MoE with strong reasoning capabilities (128K context) - \$0.10/\$0.50 per 1M tokens
@@ -31,13 +39,6 @@ https://www.baseten.co/resources/changelog/model-api-deprecation-notice-kimi-k2-
 -   `deepseek-ai/DeepSeek-V3.1` - Hybrid reasoning with advanced tool calling (163K context) - \$0.50/\$1.50 per 1M tokens
 -   `deepseek-ai/DeepSeek-V3-0324` - Fast general-purpose with enhanced reasoning (163K context) - \$0.77/\$0.77 per 1M tokens
 
-### Configuration in Cline
-
-1.  **Open Cline Settings:** Click the settings icon (⚙️) in the Cline panel.
-2.  **Select Provider:** Choose "Baseten" from the "API Provider" dropdown.
-3.  **Enter API Key:** Paste your Baseten API key into the "Baseten API Key" field.
-4.  **Select Model:** Choose your desired model from the "Model" dropdown.
-
 ### Production-First Architecture
 
 Baseten's Model APIs are built for production environments with several key advantages:
@@ -59,47 +60,17 @@ Baseten's Model APIs are built for production environments with several key adva
 
 #### Developer Experience
 - **OpenAI compatible API** - migrate by swapping a single URL
-- **Drop-in replacement** for closed models with comprehensive observability
+- **Drop-in replacement** for closed models with comprehensive observability and analytics
 - **Seamless scaling** from Model APIs to dedicated deployments
 
 ### Special Features
 
 #### Function Calling & Tool Use
-All Baseten models support structured outputs, function calling, and tool use as part of the Baseten Inference Stack, making them ideal for agentic applications.
-
-#### Reasoning Capabilities
-DeepSeek models offer enhanced reasoning with step-by-step thought processes, while maintaining production-ready performance.
-
-#### Long Context Support
-- **Up to 1 million tokens** for Llama 4 models (Maverick and Scout)
-- **262K tokens** for Qwen3 models
-- **163K tokens** for DeepSeek models
-- **Perfect for code repositories** and complex multi-turn conversations
-
-#### Quantization Optimizations
-Models are deployed with advanced quantization techniques (fp4, fp8, fp16) for optimal performance while maintaining quality.
-
-### Migration from Other Providers
-
-Baseten's OpenAI compatibility makes migration straightforward:
-
-**From OpenAI:**
-- Swap `api.openai.com` with `inference.baseten.co/v1`
-- Keep existing request/response formats
-- Benefit from significant cost savings
-
-**From Other Providers:**
-- Use standard OpenAI SDK format
-- Maintain existing prompting strategies
-- Access to newer open-source models
+All Baseten models support structured outputs, function calling, and tool use as part of the Baseten Inference Stack, making them ideal for agentic applications and coding workflows.
 
 ### Tips and Notes
 
--   **Model Selection:** Choose models based on your specific use case - reasoning models for complex tasks, coding models for development work, and flagship models for general applications.
--   **Cost Optimization:** Baseten offers some of the most competitive pricing in the market, especially for open-source models.
--   **Context Windows:** Take advantage of large context windows (up to 1M tokens) for including substantial codebases and documentation.
--   **Enterprise Ready:** Baseten is designed for production use with enterprise-grade security, compliance, and reliability.
--   **Dynamic Model Updates:** Cline automatically fetches the latest model list from Baseten, ensuring access to new models as they're released.
+-   **Dynamic Model Updates:** Cline automatically fetches the latest model list from Baseten, ensuring access to new models as they're released in real time.
 -   **Multi-Cloud Capacity Management (MCM):** Baseten's multi-cloud infrastructure ensures high availability and low latency globally.
 -   **Support:** Baseten provides dedicated support for production deployments and can work with you on dedicated resources as you scale.
 

docs/provider-config/cerebras.mdx

@@ -18,8 +18,7 @@ Cerebras delivers the world's fastest AI inference through their revolutionary w
 
 Cline supports the following Cerebras models:
 
--   `qwen-3-coder-480b-free` (Free tier) - High-performance coding model at no cost
--   `qwen-3-coder-480b` - Flagship 480B parameter coding model
+-   `zai-glm-4.6` - Intelligent general purpose model with 1,500 tokens/s
 -   `qwen-3-235b-a22b-instruct-2507` - Advanced instruction-following model
 -   `qwen-3-235b-a22b-thinking-2507` - Reasoning model with step-by-step thinking
 -   `llama-3.3-70b` - Meta's Llama 3.3 model optimized for speed

evals/package.json

@@ -1,47 +1,48 @@
 {
-  "name": "cline-evals",
-  "version": "0.1.0",
-  "description": "Evaluation scripts and tools for Cline",
-  "main": "cli/dist/index.js",
-  "scripts": {
-    "build:cli": "cd cli && tsc",
-    "start:cli": "cd cli && node dist/index.js",
-    "dev:cli": "cd cli && ts-node src/index.ts",
-    "diff-eval": "./diff-edits/run_and_open_dashboard.sh",
-    "test": "echo \"Error: no test specified\" && exit 1"
-  },
-  "keywords": [
-    "cline",
-    "evaluation",
-    "benchmark",
-    "diff-edits"
-  ],
-  "author": "",
-  "license": "MIT",
-  "dependencies": {
-    "axios": "^1.12.0",
-    "better-sqlite3": "^12.4.1",
-    "chalk": "5.6.2",
-    "dotenv": "^16.5.0",
-    "commander": "^9.4.1",
-    "execa": "^5.1.1",
-    "node-fetch": "^2.7.0",
-    "ora": "^5.4.1",
-    "sqlite": "^4.1.2",
-    "tiktoken": "^1.0.21",
-    "uuid": "^9.0.0",
-    "yargs": "^17.6.2"
-  },
-  "devDependencies": {
-    "@types/better-sqlite3": "^7.6.3",
-    "@types/node": "^18.11.18",
-    "@types/node-fetch": "^2.6.12",
-    "@types/uuid": "^9.0.0",
-    "@types/yargs": "^17.0.19",
-    "ts-node": "^10.9.1",
-    "typescript": "^4.9.4"
-  },
-  "overrides": {
-    "tar-fs": "^3.1.1"
-  }
+	"name": "cline-evals",
+	"version": "0.1.0",
+	"description": "Evaluation scripts and tools for Cline",
+	"main": "cli/dist/index.js",
+	"scripts": {
+		"build:cli": "cd cli && tsc",
+		"start:cli": "cd cli && node dist/index.js",
+		"dev:cli": "cd cli && ts-node src/index.ts",
+		"diff-eval": "./diff-edits/run_and_open_dashboard.sh",
+		"test": "echo \"Error: no test specified\" && exit 1"
+	},
+	"keywords": [
+		"cline",
+		"evaluation",
+		"benchmark",
+		"diff-edits"
+	],
+	"author": "",
+	"license": "MIT",
+	"dependencies": {
+		"axios": "^1.12.0",
+		"better-sqlite3": "^12.4.1",
+		"chalk": "5.6.2",
+		"dotenv": "^16.5.0",
+		"commander": "^9.4.1",
+		"execa": "^5.1.1",
+		"node-fetch": "^2.7.0",
+		"ora": "^5.4.1",
+		"sqlite": "^4.1.2",
+		"tiktoken": "^1.0.21",
+		"uuid": "^9.0.0",
+		"yargs": "^17.6.2"
+	},
+	"devDependencies": {
+		"@types/better-sqlite3": "^7.6.3",
+		"@types/node": "^18.11.18",
+		"@types/node-fetch": "^2.6.12",
+		"@types/uuid": "^9.0.0",
+		"@types/yargs": "^17.0.19",
+		"ts-node": "^10.9.1",
+		"typescript": "^4.9.4"
+	},
+	"overrides": {
+		"tar-fs": "^3.1.1",
+		"js-yaml": "^4.1.1"
+	}
 }

package.json

@@ -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": "3.36.1",
+	"version": "3.38.1",
 	"icon": "assets/icons/icon.png",
 	"engines": {
 		"vscode": "^1.84.0"
@@ -46,6 +46,15 @@
 	],
 	"main": "./dist/extension.js",
 	"contributes": {
+		"icons": {
+			"cline-icon": {
+				"description": "cline",
+				"default": {
+					"fontPath": "assets/icons/cline-bot.woff",
+					"fontCharacter": "\\e900"
+				}
+			}
+		},
 		"walkthroughs": [
 			{
 				"id": "ClineWalkthrough",
@@ -149,6 +158,12 @@
 				"category": "Cline",
 				"when": "cline.isDevMode"
 			},
+			{
+				"command": "cline.dev.expireMcpOAuthTokens",
+				"title": "Expire MCP OAuth Tokens (for testing)",
+				"category": "Cline",
+				"when": "cline.isDevMode"
+			},
 			{
 				"command": "cline.addToChat",
 				"title": "Add to Cline",
@@ -168,10 +183,7 @@
 				"command": "cline.generateGitCommitMessage",
 				"title": "Generate Commit Message with Cline",
 				"category": "Cline",
-				"icon": {
-					"light": "assets/icons/robot_panel_light.png",
-					"dark": "assets/icons/robot_panel_dark.png"
-				}
+				"icon": "$(cline-icon)"
 			},
 			{
 				"command": "cline.abortGitCommitMessage",
@@ -278,11 +290,11 @@
 			"commandPalette": [
 				{
 					"command": "cline.generateGitCommitMessage",
-					"when": "config.git.enabled && scmProvider == git && !cline.isGeneratingCommit"
+					"when": "config.git.enabled && !cline.isGeneratingCommit"
 				},
 				{
 					"command": "cline.abortGitCommitMessage",
-					"when": "config.git.enabled && scmProvider == git && cline.isGeneratingCommit"
+					"when": "config.git.enabled && cline.isGeneratingCommit"
 				}
 			]
 		},
@@ -331,7 +343,7 @@
 		"pretest": "npm run compile && npm run compile-tests && npm run compile-standalone && npm run lint",
 		"test": "npm-run-all test:unit test:integration",
 		"test:integration": "vscode-test",
-		"test:unit": "cross-env TS_NODE_PROJECT=./tsconfig.unit-test.json mocha",
+		"test:unit": "cross-env TS_NODE_PROJECT=./tsconfig.unit-test.json mocha # Use `UPDATE_SNAPSHOTS=true npm run test:unit` to rebuild prompt snapshots",
 		"test:coverage": "vscode-test --coverage",
 		"test:sca-server": "npx tsx watch scripts/test-standalone-core-api-server.ts",
 		"test:tp-orchestrator": "npx tsx scripts/testing-platform-orchestrator.ts",
@@ -353,7 +365,8 @@
 		"docs": "cd docs && npm run dev",
 		"docs:check-links": "cd docs && npm run check",
 		"docs:rename-file": "cd docs && npm run rename",
-		"report-issue": "node scripts/report-issue.js"
+		"report-issue": "node scripts/report-issue.js",
+		"storybook": "cd webview-ui && npm run storybook"
 	},
 	"lint-staged": {
 		"*": [
@@ -411,7 +424,7 @@
 		"@bufbuild/protobuf": "^2.2.5",
 		"@cerebras/cerebras_cloud_sdk": "^1.35.0",
 		"@google-cloud/vertexai": "^1.9.3",
-		"@google/genai": "^1.11.0",
+		"@google/genai": "^1.30.0",
 		"@grpc/grpc-js": "^1.9.15",
 		"@grpc/reflection": "^1.0.4",
 		"@mistralai/mistralai": "^1.5.0",
@@ -436,8 +449,8 @@
 		"@opentelemetry/sdk-trace-node": "^1.30.1",
 		"@opentelemetry/semantic-conventions": "^1.37.0",
 		"@playwright/test": "^1.55.1",
-		"@sap-ai-sdk/ai-api": "^1.17.0",
-		"@sap-ai-sdk/orchestration": "^1.17.0",
+		"@sap-ai-sdk/ai-api": "^2.1.0",
+		"@sap-ai-sdk/orchestration": "^2.1.0",
 		"@sentry/browser": "^9.12.0",
 		"@streamparser/json": "^0.0.22",
 		"@tailwindcss/vite": "^4.1.14",
@@ -455,7 +468,6 @@
 		"exceljs": "^4.4.0",
 		"execa": "^9.5.2",
 		"fast-deep-equal": "^3.1.3",
-		"firebase": "^11.2.0",
 		"fzf": "^0.5.2",
 		"get-folder-size": "^5.0.0",
 		"globby": "^14.0.2",
@@ -465,7 +477,6 @@
 		"image-size": "^2.0.2",
 		"isbinaryfile": "^5.0.2",
 		"jschardet": "^3.1.4",
-		"jwt-decode": "^4.0.0",
 		"mammoth": "^1.11.0",
 		"nanoid": "^5.1.6",
 		"nice-grpc": "^2.1.12",
@@ -473,7 +484,7 @@
 		"ollama": "^0.5.13",
 		"open": "^10.1.2",
 		"open-graph-scraper": "^6.9.0",
-		"openai": "^4.83.0",
+		"openai": "^6.9.0",
 		"os-name": "^6.0.0",
 		"p-mutex": "^1.0.0",
 		"p-timeout": "^6.1.4",
@@ -498,7 +509,10 @@
 		"zod": "^3.24.2"
 	},
 	"overrides": {
-		"tar-fs": ">=3.1.1"
+		"tar-fs": ">=3.1.1",
+		"tar": "^7.5.2",
+		"vite": "^7.1.11",
+		"js-yaml": "^4.1.1"
 	},
 	"c8": {
 		"reporter": [

proto/cline/file.proto

@@ -49,6 +49,9 @@ service FileService {
   // Toggle a Windsurf rule (enable or disable)
   rpc toggleWindsurfRule(ToggleWindsurfRuleRequest) returns (ClineRulesToggles);
 
+  // Toggle an Agents rule (enable or disable)
+  rpc toggleAgentsRule(ToggleAgentsRuleRequest) returns (ClineRulesToggles);
+
   // Refreshes all rule toggles (Cline, External, and Workflows)
   rpc refreshRules(EmptyRequest) returns (RefreshedRules);
 
@@ -66,6 +69,18 @@ service FileService {
 
   // Opens or creates a focus chain checklist markdown file for editing
   rpc openFocusChainFile(StringRequest) returns (Empty);
+
+  // Refreshes all hook toggles (discovers hooks and their enabled state)
+  rpc refreshHooks(EmptyRequest) returns (HooksToggles);
+
+  // Toggles a hook on or off via chmod +x/-x
+  rpc toggleHook(ToggleHookRequest) returns (ToggleHookResponse);
+
+  // Creates a new hook from template
+  rpc createHook(CreateHookRequest) returns (CreateHookResponse);
+
+  // Deletes an existing hook file
+  rpc deleteHook(DeleteHookRequest) returns (DeleteHookResponse);
 }
 
 // Response for refreshRules operation
@@ -74,8 +89,9 @@ message RefreshedRules {
   ClineRulesToggles local_cline_rules_toggles = 2;
   ClineRulesToggles local_cursor_rules_toggles = 3;
   ClineRulesToggles local_windsurf_rules_toggles = 4;
-  ClineRulesToggles local_workflow_toggles = 5;
-  ClineRulesToggles global_workflow_toggles = 6;
+  ClineRulesToggles local_agents_rules_toggles = 5;
+  ClineRulesToggles local_workflow_toggles = 6;
+  ClineRulesToggles global_workflow_toggles = 7;
 }
 
 // Request to toggle a Windsurf rule
@@ -85,6 +101,13 @@ message ToggleWindsurfRuleRequest {
   bool enabled = 3; // Whether to enable or disable the rule
 }
 
+// Request to toggle an Agents rule
+message ToggleAgentsRuleRequest {
+  Metadata metadata = 1;
+  string rule_path = 2; // Path to the rule file
+  bool enabled = 3; // Whether to enable or disable the rule
+}
+
 // Request to convert a list of URIs to relative paths
 message RelativePathsRequest {
   Metadata metadata = 1;
@@ -156,10 +179,17 @@ message RuleFile {
   bool already_exists = 3; // For createRuleFile, indicates if file already existed
 }
 
+// Enum for rule scope (local, global, or remote)
+enum RuleScope {
+  LOCAL = 0;
+  GLOBAL = 1;
+  REMOTE = 2;
+}
+
 // Request to toggle a Cline rule
 message ToggleClineRuleRequest {
   Metadata metadata = 1;
-  bool is_global = 2; // Whether this is a global rule or workspace rule
+  RuleScope scope = 2; // Scope of the rule (local, global, or remote)
   string rule_path = 3; // Path to the rule file
   bool enabled = 4; // Whether to enable or disable the rule
 }
@@ -173,6 +203,7 @@ message ClineRulesToggles {
 message ToggleClineRules {
   ClineRulesToggles global_cline_rules_toggles = 1;
   ClineRulesToggles local_cline_rules_toggles = 2;
+  ClineRulesToggles remote_rules_toggles = 3;
 }
 
 // Request to toggle a Cursor rule
@@ -187,5 +218,63 @@ message ToggleWorkflowRequest {
   Metadata metadata = 1;
   string workflow_path = 2;
   bool enabled = 3;
-  bool is_global = 4;
+  RuleScope scope = 4; // Scope of the workflow (local, global, or remote)
+}
+
+// Maps from hook name to enabled/disabled status
+message HookInfo {
+  string name = 1;
+  bool enabled = 2;
+  string absolutePath = 3;
+}
+
+message WorkspaceHooks {
+  string workspace_name = 1;
+  repeated HookInfo hooks = 2;
+}
+
+message HooksToggles {
+  repeated HookInfo global_hooks = 1;
+  repeated WorkspaceHooks workspace_hooks = 2;
+  bool is_windows = 3; // Whether the system is Windows (toggles disabled)
+}
+
+// Request to toggle a hook
+message ToggleHookRequest {
+  Metadata metadata = 1;
+  string hook_name = 2; // Name of the hook (e.g., "TaskStart")
+  bool is_global = 3; // Whether this is a global or workspace hook
+  bool enabled = 4; // Whether to enable (chmod +x) or disable (chmod -x)
+  optional string workspace_name = 5; // For multi-root workspaces, specifies which workspace
+}
+
+// Response for toggleHook operation
+message ToggleHookResponse {
+  HooksToggles hooks_toggles = 1;
+}
+
+// Request to create a hook
+message CreateHookRequest {
+  Metadata metadata = 1;
+  string hook_name = 2; // Name of the hook to create
+  bool is_global = 3; // Whether to create in global or workspace hooks directory
+  optional string workspace_name = 4; // For multi-root workspaces, specifies which workspace
+}
+
+// Response for createHook operation
+message CreateHookResponse {
+  HooksToggles hooks_toggles = 1;
+}
+
+// Request to delete a hook
+message DeleteHookRequest {
+  Metadata metadata = 1;
+  string hook_name = 2; // Name of the hook to delete
+  bool is_global = 3; // Whether this is a global or workspace hook
+  optional string workspace_name = 4; // For multi-root workspaces, specifies which workspace
+}
+
+// Response for deleteHook operation
+message DeleteHookResponse {
+  HooksToggles hooks_toggles = 1;
 }

proto/cline/mcp.proto

@@ -18,6 +18,7 @@ service McpService {
   rpc toggleToolAutoApprove(ToggleToolAutoApproveRequest) returns (McpServers);
   rpc refreshMcpMarketplace(EmptyRequest) returns (McpMarketplaceCatalog);
   rpc openMcpSettings(EmptyRequest) returns (Empty);
+  rpc authenticateMcpServer(StringRequest) returns (Empty);
 
   // Subscribe to MCP marketplace catalog updates
   rpc subscribeToMcpMarketplaceCatalog(EmptyRequest) returns (stream McpMarketplaceCatalog);
@@ -43,6 +44,7 @@ message AddRemoteMcpServerRequest {
   Metadata metadata = 1;
   string server_name = 2;
   string server_url = 3;
+  optional string transport_type = 4;
 }
 
 message ToggleToolAutoApproveRequest {
@@ -91,6 +93,8 @@ message McpServer {
   repeated McpResourceTemplate resource_templates = 7;
   optional bool disabled = 8;
   optional int32 timeout = 9;
+  optional bool oauth_required = 10;
+  optional string oauth_auth_status = 11;
 }
 
 message McpServers {

proto/cline/models.proto

@@ -97,6 +97,7 @@ message OpenRouterModelInfo {
   optional bool supports_global_endpoint = 11;
   repeated ModelTier tiers = 12;
   optional string name = 13;
+  optional double temperature = 14;
 }
 
 // Shared response message for model information
@@ -422,6 +423,7 @@ enum ApiProvider {
   MINIMAX = 36;
   HICAP = 37;
   AIHUBMIX = 38;
+  NOUSRESEARCH = 39;
 }
 
 // Model info for OpenAI-compatible models
@@ -546,6 +548,7 @@ message ModelsApiConfiguration {
   optional string aihubmix_api_key = 82;
   optional string aihubmix_base_url = 83;
   optional string aihubmix_app_code = 84;
+  optional string nous_research_api_key = 85;
 
   // Plan mode configurations
   optional ApiProvider plan_mode_api_provider = 100;
@@ -585,6 +588,8 @@ message ModelsApiConfiguration {
   optional OpenRouterModelInfo plan_mode_hicap_model_info = 134;
   optional string plan_mode_aihubmix_model_id = 135;
   optional OpenAiCompatibleModelInfo plan_mode_aihubmix_model_info = 136;
+  optional string plan_mode_nous_research_model_id = 137;
+  optional string gemini_plan_mode_thinking_level = 138;
 
   // Act mode configurations
   optional ApiProvider act_mode_api_provider = 200;
@@ -624,4 +629,6 @@ message ModelsApiConfiguration {
   optional OpenRouterModelInfo act_mode_hicap_model_info = 234;
   optional string act_mode_aihubmix_model_id = 235;
   optional OpenAiCompatibleModelInfo act_mode_aihubmix_model_info = 236;
+  optional string act_mode_nous_research_model_id = 237;
+  optional string gemini_act_mode_thinking_level = 238;
 }

proto/cline/state.proto

@@ -31,6 +31,7 @@ service StateService {
   rpc installClineCli(EmptyRequest) returns (Empty);
   rpc checkCliInstallation(EmptyRequest) returns (Boolean);
   rpc getProcessInfo(EmptyRequest) returns (ProcessInfo);
+  rpc flushPendingState(EmptyRequest) returns (Empty);
 }
 
 message AutoApprovalActions {
@@ -89,6 +90,7 @@ message Secrets {
   optional string oca_api_key = 37;
   optional string oca_refresh_token = 38;
   optional string hicap_api_key = 39;
+  optional string mcp_oauth_secrets = 40;
 }
 
 message Settings {
@@ -360,6 +362,7 @@ message UpdateSettingsRequest {
   optional int32 subagent_terminal_output_line_limit = 30;
   optional string cline_env = 31;
   optional bool native_tool_call_enabled = 32;
+  optional OnboardingModelGroup onboarding_models = 33;
 }
 
 message UpdateTerminalConnectionTimeoutRequest {
@@ -387,3 +390,17 @@ message OnboardingProgressRequest {
   optional bool completed = 3;
   optional string model_selected = 4;
 }
+
+message OnboardingModelGroup {
+  repeated OnboardingModel models = 1;
+}
+
+message OnboardingModel {
+  string id = 1;
+  string name = 2;
+  int32 score = 3;
+  int32 latency = 4;
+  string badge = 5;
+  string group = 6;
+  OpenRouterModelInfo info = 7;
+}

proto/cline/task.proto

@@ -70,6 +70,7 @@ message TaskResponse {
   int32 tokens_out = 8;
   int32 cache_writes = 9;
   int32 cache_reads = 10;
+  string model_id = 11;
 }
 
 // Request for getting task history with filtering
@@ -99,6 +100,7 @@ message TaskItem {
   int32 tokens_out = 8;
   int32 cache_writes = 9;
   int32 cache_reads = 10;
+  string model_id = 11;
 }
 
 // Request for ask response operation

proto/cline/ui.proto

@@ -32,6 +32,7 @@ enum ClineAsk {
   CONDENSE = 13;
   REPORT_BUG = 14;
   SUMMARIZE_TASK = 15;
+  ACT_MODE_RESPOND = 16;
 }
 
 // Enum for ClineSay types
@@ -184,6 +185,11 @@ message ClineApiReqInfo {
   ApiReqRetryStatus retry_status = 9;
 }
 
+message ClineModelInfo {
+  string provider_id = 1;
+  string model_id = 2;
+}
+
 // Main ClineMessage type
 message ClineMessage {
   int64 ts = 1;
@@ -210,6 +216,7 @@ message ClineMessage {
   ClineAskQuestion ask_question = 20;
   ClineAskNewTask ask_new_task = 21;
   ClineApiReqInfo api_req_info = 22;
+  ClineModelInfo model_info = 23;
 }
 
 // UiService provides methods for managing UI interactions

scripts/cli-providers.mjs

@@ -95,6 +95,7 @@ const ENABLED_PROVIDERS = [
 	"ollama", // Ollama local models
 	"cerebras", // Cerebras models
 	"oca", // Oracle Code Assist
+	"nousResearch", // NousResearch provider
 ]
 
 /**

scripts/download-ripgrep.mjs

@@ -11,7 +11,7 @@ import fs from "fs"
 import https from "https"
 import path from "path"
 import { pipeline } from "stream/promises"
-import tar from "tar"
+import * as tar from "tar"
 import { promisify } from "util"
 import { createGunzip } from "zlib"
 

src/api/providers/dify.ts

@@ -1,5 +1,5 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ModelInfo } from "@shared/api"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler } from "../../core/api/index"
 import { ApiStream } from "../../core/api/transform/stream"
@@ -33,7 +33,7 @@ export class DifyHandler implements ApiHandler {
 		}
 	}
 
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		console.log("[DIFY DEBUG] createMessage called with:", {
 			systemPromptLength: systemPrompt?.length || 0,
 			messagesCount: messages?.length || 0,
@@ -255,7 +255,7 @@ export class DifyHandler implements ApiHandler {
 		}
 	}
 
-	private convertMessagesToQuery(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): string {
+	private convertMessagesToQuery(systemPrompt: string, messages: ClineStorageMessage[]): string {
 		// Dify's context is managed by `conversation_id`. The `query` should be the last user message.
 		// The system prompt is typically configured in the Dify App itself.
 		const lastUserMessage = messages.filter((m) => m.role === "user").pop()

src/config.ts

@@ -9,14 +9,6 @@ export interface EnvironmentConfig {
 	appBaseUrl: string
 	apiBaseUrl: string
 	mcpBaseUrl: string
-	firebase: {
-		apiKey: string
-		authDomain: string
-		projectId: string
-		storageBucket?: string
-		messagingSenderId?: string
-		appId?: string
-	}
 }
 
 class ClineEndpoint {
@@ -28,8 +20,8 @@ class ClineEndpoint {
 	private environment: Environment = Environment.production
 
 	private constructor() {
-		// Set environment at module load
-		const _env = process?.env?.CLINE_ENVIRONMENT
+		// Set environment at module load. Use override if provided.
+		const _env = process?.env?.CLINE_ENVIRONMENT_OVERRIDE || process?.env?.CLINE_ENVIRONMENT
 		if (_env && Object.values(Environment).includes(_env as Environment)) {
 			this.environment = _env as Environment
 			return
@@ -63,14 +55,6 @@ class ClineEndpoint {
 					appBaseUrl: "https://staging-app.cline.bot",
 					apiBaseUrl: "https://core-api.staging.int.cline.bot",
 					mcpBaseUrl: "https://core-api.staging.int.cline.bot/v1/mcp",
-					firebase: {
-						apiKey: "AIzaSyASSwkwX1kSO8vddjZkE5N19QU9cVQ0CIk",
-						authDomain: "cline-staging.firebaseapp.com",
-						projectId: "cline-staging",
-						storageBucket: "cline-staging.firebasestorage.app",
-						messagingSenderId: "853479478430",
-						appId: "1:853479478430:web:2de0dba1c63c3262d4578f",
-					},
 				}
 			case Environment.local:
 				return {
@@ -78,11 +62,6 @@ class ClineEndpoint {
 					appBaseUrl: "http://localhost:3000",
 					apiBaseUrl: "http://localhost:7777",
 					mcpBaseUrl: "https://api.cline.bot/v1/mcp",
-					firebase: {
-						apiKey: "AIzaSyD8wtkd1I-EICuAg6xgAQpRdwYTvwxZG2w",
-						authDomain: "cline-preview.firebaseapp.com",
-						projectId: "cline-preview",
-					},
 				}
 			default:
 				return {
@@ -90,14 +69,6 @@ class ClineEndpoint {
 					appBaseUrl: "https://app.cline.bot",
 					apiBaseUrl: "https://api.cline.bot",
 					mcpBaseUrl: "https://api.cline.bot/v1/mcp",
-					firebase: {
-						apiKey: "AIzaSyC5rx59Xt8UgwdU3PCfzUF7vCwmp9-K2vk",
-						authDomain: "cline-prod.firebaseapp.com",
-						projectId: "cline-prod",
-						storageBucket: "cline-prod.firebasestorage.app",
-						messagingSenderId: "941048379330",
-						appId: "1:941048379330:web:45058eedeefc5cdfcc485b",
-					},
 				}
 		}
 	}

src/core/api/index.ts

@@ -1,6 +1,6 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ApiConfiguration, ModelInfo, QwenApiRegions } from "@shared/api"
 import { Mode } from "@shared/storage/types"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { ClineTool } from "@/shared/tools"
 import { AIhubmixHandler } from "./providers/aihubmix"
 import { AnthropicHandler } from "./providers/anthropic"
@@ -25,6 +25,7 @@ import { MinimaxHandler } from "./providers/minimax"
 import { MistralHandler } from "./providers/mistral"
 import { MoonshotHandler } from "./providers/moonshot"
 import { NebiusHandler } from "./providers/nebius"
+import { NousResearchHandler } from "./providers/nousresearch"
 import { OcaHandler } from "./providers/oca"
 import { OllamaHandler } from "./providers/ollama"
 import { OpenAiHandler } from "./providers/openai"
@@ -46,9 +47,8 @@ import { ApiStream, ApiStreamUsageChunk } from "./transform/stream"
 export type CommonApiHandlerOptions = {
 	onRetryAttempt?: ApiConfiguration["onRetryAttempt"]
 }
-
 export interface ApiHandler {
-	createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: ClineTool[]): ApiStream
+	createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: ClineTool[], useResponseApi?: boolean): ApiStream
 	getModel(): ApiHandlerModel
 	getApiStreamUsage?(): Promise<ApiStreamUsageChunk | undefined>
 }
@@ -94,6 +94,7 @@ function createHandlerForProvider(
 				reasoningEffort: mode === "plan" ? options.planModeReasoningEffort : options.actModeReasoningEffort,
 				thinkingBudgetTokens:
 					mode === "plan" ? options.planModeThinkingBudgetTokens : options.actModeThinkingBudgetTokens,
+				geminiThinkingLevel: mode === "plan" ? options.geminiPlanModeThinkingLevel : options.geminiActModeThinkingLevel,
 			})
 		case "bedrock":
 			return new AwsBedrockHandler({
@@ -166,6 +167,7 @@ function createHandlerForProvider(
 				geminiBaseUrl: options.geminiBaseUrl,
 				thinkingBudgetTokens:
 					mode === "plan" ? options.planModeThinkingBudgetTokens : options.actModeThinkingBudgetTokens,
+				thinkingLevel: mode === "plan" ? options.geminiPlanModeThinkingLevel : options.geminiActModeThinkingLevel,
 				apiModelId: mode === "plan" ? options.planModeApiModelId : options.actModeApiModelId,
 				ulid: options.ulid,
 			})
@@ -250,6 +252,7 @@ function createHandlerForProvider(
 				openRouterProviderSorting: options.openRouterProviderSorting,
 				openRouterModelId: mode === "plan" ? options.planModeOpenRouterModelId : options.actModeOpenRouterModelId,
 				openRouterModelInfo: mode === "plan" ? options.planModeOpenRouterModelInfo : options.actModeOpenRouterModelInfo,
+				geminiThinkingLevel: mode === "plan" ? options.geminiPlanModeThinkingLevel : options.geminiActModeThinkingLevel,
 			})
 		case "litellm":
 			return new LiteLlmHandler({
@@ -416,6 +419,12 @@ function createHandlerForProvider(
 				hicapApiKey: options.hicapApiKey,
 				hicapModelId: mode === "plan" ? options.planModeHicapModelId : options.actModeHicapModelId,
 			})
+		case "nousResearch":
+			return new NousResearchHandler({
+				onRetryAttempt: options.onRetryAttempt,
+				nousResearchApiKey: options.nousResearchApiKey,
+				apiModelId: mode === "plan" ? options.planModeNousResearchModelId : options.actModeNousResearchModelId,
+			})
 		default:
 			return new AnthropicHandler({
 				onRetryAttempt: options.onRetryAttempt,

src/core/api/providers/__tests__/claude-code.test.ts

@@ -1,8 +1,8 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { afterEach, beforeEach, describe, it } from "mocha"
 import sinon from "sinon"
 import "should"
 import { ClaudeCodeHandler } from "@core/api/providers/claude-code"
+import { ClineStorageMessage } from "@/shared/messages/content"
 
 describe("ClaudeCodeHandler", () => {
 	let handler: ClaudeCodeHandler
@@ -71,7 +71,7 @@ describe("ClaudeCodeHandler", () => {
 			runClaudeCodeStub.returns(mockGenerator() as any)
 
 			const systemPrompt = "You are a helpful assistant."
-			const messages: Anthropic.Messages.MessageParam[] = [{ role: "user", content: "Hello" }]
+			const messages: ClineStorageMessage[] = [{ role: "user", content: "Hello" }]
 
 			const usageData: any[] = []
 
@@ -140,7 +140,7 @@ describe("ClaudeCodeHandler", () => {
 			runClaudeCodeStub.returns(mockGenerator() as any)
 
 			const systemPrompt = "You are a helpful assistant."
-			const messages: Anthropic.Messages.MessageParam[] = [{ role: "user", content: "Hello" }]
+			const messages: ClineStorageMessage[] = [{ role: "user", content: "Hello" }]
 
 			const usageData: any[] = []
 
@@ -199,7 +199,7 @@ describe("ClaudeCodeHandler", () => {
 			runClaudeCodeStub.returns(mockGenerator() as any)
 
 			const systemPrompt = "You are a helpful assistant."
-			const messages: Anthropic.Messages.MessageParam[] = [{ role: "user", content: "Hello" }]
+			const messages: ClineStorageMessage[] = [{ role: "user", content: "Hello" }]
 
 			const usageData: any[] = []
 

src/core/api/providers/__tests__/litellm.test.ts

@@ -1,8 +1,8 @@
-import Anthropic from "@anthropic-ai/sdk"
 import { LiteLlmHandler, type LiteLlmModelInfoResponse } from "@core/api/providers/litellm"
 import { convertToOpenAiMessages } from "@core/api/transform/openai-format"
 import { expect } from "chai"
 import sinon from "sinon"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { mockFetchForTesting } from "@/shared/net"
 
 const fakeClient = {
@@ -109,7 +109,7 @@ describe("LiteLlmHandler", () => {
 
 			it("sends the system prompt and messages with the openai format", async () => {
 				const systemPrompt = "Test System Prompt"
-				const messages: Anthropic.Messages.MessageParam[] = [
+				const messages: ClineStorageMessage[] = [
 					{
 						role: "user",
 						content: "first message",
@@ -161,7 +161,7 @@ describe("LiteLlmHandler", () => {
 
 			it("inserts the cache control in the system prompt and the last two user messages", async () => {
 				const systemPrompt = "Test System Prompt"
-				const messages: Anthropic.Messages.MessageParam[] = [
+				const messages: ClineStorageMessage[] = [
 					{
 						role: "user",
 						content: "first message",

src/core/api/providers/__tests__/ollama.test.ts

@@ -1,9 +1,9 @@
 import { afterEach, before, beforeEach, describe, it } from "mocha"
 import "should"
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ApiHandlerOptions } from "@shared/api"
 import axios from "axios"
 import sinon from "sinon"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { OllamaHandler } from "../ollama"
 
 describe("OllamaHandler", () => {
@@ -59,7 +59,7 @@ describe("OllamaHandler", () => {
 			} as any)
 
 			const systemPrompt = "You are a helpful assistant."
-			const messages: Anthropic.Messages.MessageParam[] = [{ role: "user", content: "Hello" }]
+			const messages: ClineStorageMessage[] = [{ role: "user", content: "Hello" }]
 
 			const result = []
 			const usageInfo = []
@@ -114,7 +114,7 @@ describe("OllamaHandler", () => {
 			}
 
 			const systemPrompt = "You are a helpful assistant."
-			const messages: Anthropic.Messages.MessageParam[] = [{ role: "user", content: "Hello" }]
+			const messages: ClineStorageMessage[] = [{ role: "user", content: "Hello" }]
 
 			// Start the request and catch the error
 			let errorMessage = ""
@@ -158,7 +158,7 @@ describe("OllamaHandler", () => {
 			} as any)
 
 			const systemPrompt = "You are a helpful assistant."
-			const messages: Anthropic.Messages.MessageParam[] = [{ role: "user", content: "Hello" }]
+			const messages: ClineStorageMessage[] = [{ role: "user", content: "Hello" }]
 
 			const result = []
 
@@ -204,7 +204,7 @@ describe("OllamaHandler", () => {
 			}
 
 			const systemPrompt = "You are a helpful assistant."
-			const messages: Anthropic.Messages.MessageParam[] = [{ role: "user", content: "Hello" }]
+			const messages: ClineStorageMessage[] = [{ role: "user", content: "Hello" }]
 
 			const result = []
 

src/core/api/providers/anthropic.ts

@@ -2,8 +2,8 @@ import { Anthropic } from "@anthropic-ai/sdk"
 import { Tool as AnthropicTool } from "@anthropic-ai/sdk/resources/index"
 import { Stream as AnthropicStream } from "@anthropic-ai/sdk/streaming"
 import { AnthropicModelId, anthropicDefaultModelId, anthropicModels, CLAUDE_SONNET_1M_SUFFIX, ModelInfo } from "@shared/api"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
-import { ClineTool } from "@/shared/tools"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
 import { withRetry } from "../retry"
 import { sanitizeAnthropicMessages } from "../transform/anthropic-format"
@@ -43,7 +43,7 @@ export class AnthropicHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: ClineTool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: AnthropicTool[]): ApiStream {
 		const client = this.ensureClient()
 
 		const model = this.getModel()
@@ -74,19 +74,7 @@ export class AnthropicHandler implements ApiHandler {
 			case "claude-opus-4-1-20250805":
 			case "claude-3-opus-20240229":
 			case "claude-3-haiku-20240307": {
-				/*
-				The latest message will be the new user message, one before will be the assistant message from a previous request, and the user message before that will be a previously cached user message. So we need to mark the latest user message as ephemeral to cache it for the next request, and mark the second to last user message as ephemeral to let the server know the last message to retrieve from the cache for the current request..
-				*/
-				const userMsgIndices = messages.reduce((acc, msg, index) => {
-					if (msg.role === "user") {
-						acc.push(index)
-					}
-					return acc
-				}, [] as number[])
-				const lastUserMsgIndex = userMsgIndices[userMsgIndices.length - 1] ?? -1
-				const secondLastMsgUserIndex = userMsgIndices[userMsgIndices.length - 2] ?? -1
-
-				const anthropicMessages = sanitizeAnthropicMessages(messages, lastUserMsgIndex, secondLastMsgUserIndex)
+				const anthropicMessages = sanitizeAnthropicMessages(messages, true)
 
 				stream = await client.messages.create(
 					{
@@ -106,7 +94,7 @@ export class AnthropicHandler implements ApiHandler {
 						messages: anthropicMessages,
 						// tools, // cache breakpoints go from tools > system > messages, and since tools dont change, we can just set the breakpoint at the end of system (this avoids having to set a breakpoint at the end of tools which by itself does not meet min requirements for haiku caching)
 						stream: true,
-						tools: nativeToolsOn ? (tools as AnthropicTool[]) : undefined,
+						tools: nativeToolsOn ? tools : undefined,
 						// tool_choice options:
 						// - none: disables tool use, even if tools are provided. Claude will not call any tools.
 						// - auto: allows Claude to decide whether to call any provided tools or not. This is the default value when tools are provided.
@@ -135,29 +123,30 @@ export class AnthropicHandler implements ApiHandler {
 					max_tokens: model.info.maxTokens || 8192,
 					temperature: 0,
 					system: [{ text: systemPrompt, type: "text" }],
-					messages: sanitizeAnthropicMessages(messages),
-					// tools,
-					// tool_choice: { type: "auto" },
+					messages: sanitizeAnthropicMessages(messages, false),
+					tools: nativeToolsOn ? tools : undefined,
+					tool_choice: { type: "auto" },
 					stream: true,
 				})
 				break
 			}
 		}
 
-		let thinkingDeltaAccumulator = ""
 		const lastStartedToolCall = { id: "", name: "", arguments: "" }
 
 		for await (const chunk of stream) {
 			switch (chunk?.type) {
 				case "message_start":
-					// tells us cache reads/writes/input/output
-					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,
+					{
+						// tells us cache reads/writes/input/output
+						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":
@@ -178,15 +167,7 @@ export class AnthropicHandler implements ApiHandler {
 							yield {
 								type: "reasoning",
 								reasoning: chunk.content_block.thinking || "",
-							}
-							const thinking = chunk.content_block.thinking
-							const signature = chunk.content_block.signature
-							if (thinking && signature) {
-								yield {
-									type: "ant_thinking",
-									thinking,
-									signature,
-								}
+								signature: chunk.content_block.signature,
 							}
 							break
 						case "redacted_thinking":
@@ -194,10 +175,7 @@ export class AnthropicHandler implements ApiHandler {
 							yield {
 								type: "reasoning",
 								reasoning: "[Redacted thinking block]",
-							}
-							yield {
-								type: "ant_redacted_thinking",
-								data: chunk.content_block.data,
+								redacted_data: chunk.content_block.data,
 							}
 							break
 						case "tool_use":
@@ -231,15 +209,14 @@ export class AnthropicHandler implements ApiHandler {
 								type: "reasoning",
 								reasoning: chunk.delta.thinking,
 							}
-							thinkingDeltaAccumulator += chunk.delta.thinking
 							break
 						case "signature_delta":
 							// It's used when sending the thinking block back to the API
 							// API expects this in completed form, not as array of deltas
-							if (thinkingDeltaAccumulator && chunk.delta.signature) {
+							if (chunk.delta.signature) {
 								yield {
-									type: "ant_thinking",
-									thinking: thinkingDeltaAccumulator,
+									type: "reasoning",
+									reasoning: "", // reasoning text is already sent via thinking_delta
 									signature: chunk.delta.signature,
 								}
 							}

src/core/api/providers/asksage.ts

@@ -1,5 +1,5 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { AskSageModelId, askSageDefaultModelId, askSageDefaultURL, askSageModels, ModelInfo } from "@shared/api"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from ".."
 import { withRetry } from "../retry"
@@ -47,7 +47,7 @@ export class AskSageHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		try {
 			const model = this.getModel()
 

src/core/api/providers/baseten.ts

@@ -1,12 +1,14 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { BasetenModelId, basetenDefaultModelId, basetenModels, ModelInfo } from "@shared/api"
 import { calculateApiCostOpenAI } from "@utils/cost"
 import OpenAI from "openai"
+import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { ApiStream } from "../transform/stream"
+import { ToolCallProcessor } from "../transform/tool-call-processor"
 
 interface BasetenHandlerOptions extends CommonApiHandlerOptions {
 	basetenApiKey?: string
@@ -98,10 +100,11 @@ export class BasetenHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 		const maxTokens = this.getOptimalMaxTokens(model)
+		const toolCallProcessor = new ToolCallProcessor()
 
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [
 			{ role: "system", content: systemPrompt },
@@ -115,21 +118,22 @@ export class BasetenHandler implements ApiHandler {
 			stream: true,
 			stream_options: { include_usage: true },
 			temperature: 0,
+			tools,
+			tool_choice: tools && tools.length > 0 ? "auto" : undefined,
 		})
 
 		let didOutputUsage = false
 
 		for await (const chunk of stream) {
-			const delta = chunk.choices[0]?.delta
+			const delta = chunk?.choices?.[0]?.delta
 
 			// Handle reasoning field if present (for reasoning models with parsed output)
-			if ((delta as any)?.reasoning) {
-				const reasoningContent = (delta as any).reasoning as string
+			if (delta && "reasoning" in delta && delta?.reasoning) {
+				const reasoning = typeof delta.reasoning === "string" ? delta.reasoning : JSON.stringify(delta.reasoning)
 				yield {
 					type: "reasoning",
-					reasoning: reasoningContent,
+					reasoning,
 				}
-				continue
 			}
 
 			// Handle content field
@@ -140,6 +144,10 @@ export class BasetenHandler implements ApiHandler {
 				}
 			}
 
+			if (delta?.tool_calls) {
+				yield* toolCallProcessor.processToolCallDeltas(delta.tool_calls)
+			}
+
 			// Handle usage information - only output once
 			if (!didOutputUsage && chunk.usage) {
 				yield* this.yieldUsage(model.info, chunk.usage)

src/core/api/providers/bedrock.ts

@@ -1,4 +1,3 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 // Import proper AWS SDK types
 import type { ContentBlock, Message } from "@aws-sdk/client-bedrock-runtime"
 import {
@@ -12,6 +11,7 @@ import { fromNodeProviderChain } from "@aws-sdk/credential-providers"
 import { BedrockModelId, bedrockDefaultModelId, bedrockModels, CLAUDE_SONNET_1M_SUFFIX, ModelInfo } from "@shared/api"
 import { calculateApiCostOpenAI, calculateApiCostQwen } from "@utils/cost"
 import { ExtensionRegistryInfo } from "@/registry"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
 import { convertToR1Format } from "../transform/r1-format"
@@ -121,7 +121,7 @@ export class AwsBedrockHandler implements ApiHandler {
 	}
 
 	@withRetry({ maxRetries: 4 })
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		// cross region inference requires prefixing the model id with the region
 		const rawModelId = await this.getModelId()
 
@@ -342,7 +342,7 @@ export class AwsBedrockHandler implements ApiHandler {
 	 */
 	private async *createDeepseekMessage(
 		systemPrompt: string,
-		messages: Anthropic.Messages.MessageParam[],
+		messages: ClineStorageMessage[],
 		modelId: string,
 		model: { id: string; info: ModelInfo },
 	): ApiStream {
@@ -480,7 +480,7 @@ export class AwsBedrockHandler implements ApiHandler {
 	 * First uses convertToR1Format to merge consecutive messages with the same role,
 	 * then converts to the string format that DeepSeek R1 expects
 	 */
-	private formatDeepseekR1Prompt(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): string {
+	private formatDeepseekR1Prompt(systemPrompt: string, messages: ClineStorageMessage[]): string {
 		// First use convertToR1Format to merge consecutive messages with the same role
 		const r1Messages = convertToR1Format([{ role: "user", content: systemPrompt }, ...messages])
 
@@ -513,7 +513,7 @@ export class AwsBedrockHandler implements ApiHandler {
 	 * Estimates token count based on text length (approximate)
 	 * Note: This is a rough estimation, as the actual token count depends on the tokenizer
 	 */
-	private estimateInputTokens(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): number {
+	private estimateInputTokens(systemPrompt: string, messages: ClineStorageMessage[]): number {
 		// For Deepseek R1, we estimate the token count of the formatted prompt
 		// The formatted prompt includes special tokens and consistent formatting
 		const formattedPrompt = this.formatDeepseekR1Prompt(systemPrompt, messages)
@@ -680,11 +680,7 @@ export class AwsBedrockHandler implements ApiHandler {
 				}
 			}
 		} catch (error) {
-			console.error("Error processing Converse API response:", error)
-			yield {
-				type: "text",
-				text: `[ERROR] Failed to process response: ${error instanceof Error ? error.message : String(error)}`,
-			}
+			throw error
 		}
 	}
 
@@ -703,9 +699,20 @@ export class AwsBedrockHandler implements ApiHandler {
 				text: `[ERROR] Model stream error: ${chunk.modelStreamErrorException.message}`,
 			}
 		} else if (chunk.validationException) {
+			// Check if this is a context window error - if so, throw it
+			// so the retry mechanism can handle truncation
+			const message = chunk.validationException.message || ""
+			const isContextError = /input.*too long|context.*exceed|maximum.*token|input length.*max.*tokens/i.test(message)
+
+			if (isContextError) {
+				// Throw as exception so context management can handle it
+				throw chunk.validationException
+			}
+
+			// Otherwise yield as error text
 			yield {
 				type: "text",
-				text: `[ERROR] Validation error: ${chunk.validationException.message}`,
+				text: `[ERROR] Validation error: ${message}`,
 			}
 		} else if (chunk.throttlingException) {
 			yield {
@@ -779,7 +786,7 @@ export class AwsBedrockHandler implements ApiHandler {
 	 */
 	private async *createAnthropicMessage(
 		systemPrompt: string,
-		messages: Anthropic.Messages.MessageParam[],
+		messages: ClineStorageMessage[],
 		modelId: string,
 		model: { id: string; info: ModelInfo },
 		enable1mContextWindow: boolean,
@@ -835,7 +842,7 @@ export class AwsBedrockHandler implements ApiHandler {
 	 * Formats messages for models using the Converse API specification
 	 * Used by both Anthropic and Nova models to avoid code duplication
 	 */
-	private formatMessagesForConverseAPI(messages: Anthropic.Messages.MessageParam[]): Message[] {
+	private formatMessagesForConverseAPI(messages: ClineStorageMessage[]): Message[] {
 		return messages.map((message) => {
 			// Determine role (user or assistant)
 			const role = message.role === "user" ? ConversationRole.USER : ConversationRole.ASSISTANT
@@ -968,7 +975,7 @@ export class AwsBedrockHandler implements ApiHandler {
 	 */
 	private async *createNovaMessage(
 		systemPrompt: string,
-		messages: Anthropic.Messages.MessageParam[],
+		messages: ClineStorageMessage[],
 		modelId: string,
 		model: { id: string; info: ModelInfo },
 	): ApiStream {
@@ -1008,7 +1015,7 @@ export class AwsBedrockHandler implements ApiHandler {
 	 */
 	private async *createOpenAIMessage(
 		systemPrompt: string,
-		messages: Anthropic.Messages.MessageParam[],
+		messages: ClineStorageMessage[],
 		modelId: string,
 		model: { id: string; info: ModelInfo },
 	): ApiStream {
@@ -1143,7 +1150,7 @@ export class AwsBedrockHandler implements ApiHandler {
 	 */
 	private async *createQwenMessage(
 		systemPrompt: string,
-		messages: Anthropic.Messages.MessageParam[],
+		messages: ClineStorageMessage[],
 		modelId: string,
 		model: { id: string; info: ModelInfo },
 	): ApiStream {

src/core/api/providers/cerebras.ts

@@ -1,6 +1,6 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import Cerebras from "@cerebras/cerebras_cloud_sdk"
 import { CerebrasModelId, cerebrasDefaultModelId, cerebrasModels, ModelInfo } from "@shared/api"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
 import { withRetry } from "../retry"
@@ -46,7 +46,7 @@ export class CerebrasHandler implements ApiHandler {
 		baseDelay: 5000, // Start with 5 second delay
 		maxDelay: 60000, // Allow up to 60 second delays to respect rate limits
 	})
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		const client = this.ensureClient()
 
 		// Convert Anthropic messages to Cerebras format

src/core/api/providers/claude-code.ts

@@ -1,7 +1,7 @@
-import type { Anthropic } from "@anthropic-ai/sdk"
 import { filterMessagesForClaudeCode } from "@/integrations/claude-code/message-filter"
 import { runClaudeCode } from "@/integrations/claude-code/run"
 import { ClaudeCodeModelId, claudeCodeDefaultModelId, claudeCodeModels } from "@/shared/api"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { type ApiHandler, CommonApiHandlerOptions } from ".."
 import { withRetry } from "../retry"
 import { type ApiStream, ApiStreamUsageChunk } from "../transform/stream"
@@ -24,7 +24,7 @@ export class ClaudeCodeHandler implements ApiHandler {
 		baseDelay: 2000,
 		maxDelay: 15000,
 	})
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		// Filter out image blocks since Claude Code doesn't support them
 		const filteredMessages = filterMessagesForClaudeCode(messages)
 
@@ -113,7 +113,18 @@ export class ClaudeCodeHandler implements ApiHandler {
 							}
 							break
 						case "tool_use":
-							console.error(`tool_use is not supported yet. Received: ${JSON.stringify(content)}`)
+							// Yield tool_use blocks to the streaming pipeline for proper tool execution
+							yield {
+								type: "tool_calls",
+								tool_call: {
+									call_id: content.id,
+									function: {
+										id: content.id,
+										name: content.name,
+										arguments: content.input,
+									},
+								},
+							}
 							break
 					}
 				}

src/core/api/providers/cline.ts

@@ -1,4 +1,3 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ModelInfo, openRouterDefaultModelId, openRouterDefaultModelInfo } from "@shared/api"
 import { shouldSkipReasoningForModel } from "@utils/model-utils"
 import axios from "axios"
@@ -8,7 +7,9 @@ import { ClineEnv } from "@/config"
 import { ClineAccountService } from "@/services/account/ClineAccountService"
 import { AuthService } from "@/services/auth/AuthService"
 import { buildClineExtraHeaders } from "@/services/EnvUtils"
+import { Logger } from "@/services/logging/Logger"
 import { CLINE_ACCOUNT_AUTH_ERROR_MESSAGE } from "@/shared/ClineAccount"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch, getAxiosSettings } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -26,6 +27,7 @@ interface ClineHandlerOptions extends CommonApiHandlerOptions {
 	openRouterModelId?: string
 	openRouterModelInfo?: ModelInfo
 	clineAccountId?: string
+	geminiThinkingLevel?: string
 }
 
 export class ClineHandler implements ApiHandler {
@@ -96,7 +98,7 @@ export class ClineHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		try {
 			const client = await this.ensureClient()
 
@@ -114,11 +116,13 @@ export class ClineHandler implements ApiHandler {
 				this.options.thinkingBudgetTokens,
 				this.options.openRouterProviderSorting,
 				tools,
+				this.options.geminiThinkingLevel,
 			)
 
 			const toolCallProcessor = new ToolCallProcessor()
 
 			for await (const chunk of stream) {
+				Logger.debug("ClineHandler chunk:" + JSON.stringify(chunk))
 				// openrouter returns an error object instead of the openai sdk throwing an error
 				if ("error" in chunk) {
 					const error = chunk.error as OpenRouterErrorResponse["error"]
@@ -149,6 +153,7 @@ export class ClineHandler implements ApiHandler {
 				}
 
 				const delta = choice?.delta
+
 				if (delta?.content) {
 					yield {
 						type: "text",
@@ -165,8 +170,7 @@ export class ClineHandler implements ApiHandler {
 				if ("reasoning" in delta && delta.reasoning && !shouldSkipReasoningForModel(this.options.openRouterModelId)) {
 					yield {
 						type: "reasoning",
-						// @ts-ignore-next-line
-						reasoning: delta.reasoning,
+						reasoning: typeof delta.reasoning === "string" ? delta.reasoning : JSON.stringify(delta.reasoning),
 					}
 				}
 
@@ -181,20 +185,23 @@ export class ClineHandler implements ApiHandler {
 					"reasoning_details" in delta &&
 					delta.reasoning_details &&
 					// @ts-ignore-next-line
-					delta.reasoning_details.length && // exists and non-0
+					delta?.reasoning_details?.length && // exists and non-0
 					!shouldSkipReasoningForModel(this.options.openRouterModelId)
 				) {
 					yield {
-						type: "reasoning_details",
-						reasoning_details: delta.reasoning_details,
+						type: "reasoning",
+						reasoning: "",
+						details: delta.reasoning_details,
 					}
 				}
 
+				console.log("didOutputUsage", didOutputUsage, chunk.usage)
+
 				if (!didOutputUsage && chunk.usage) {
 					// @ts-ignore-next-line
 					let totalCost = (chunk.usage.cost || 0) + (chunk.usage.cost_details?.upstream_inference_cost || 0)
 
-					if (this.getModel().id === "x-ai/grok-code-fast-1") {
+					if (this.getModel().id === "x-ai/grok-code-fast-1" || this.getModel().id === "minimax/minimax-m2") {
 						totalCost = 0
 					}
 

src/core/api/providers/deepseek.ts

@@ -1,8 +1,8 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { DeepSeekModelId, deepSeekDefaultModelId, deepSeekModels, ModelInfo } from "@shared/api"
 import { calculateApiCostOpenAI } from "@utils/cost"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -75,7 +75,7 @@ export class DeepSeekHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 

src/core/api/providers/dify.ts

@@ -1,4 +1,4 @@
-import { Anthropic } from "@anthropic-ai/sdk"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ModelInfo } from "../../../shared/api"
 import { ApiHandler } from "../index"
@@ -97,7 +97,7 @@ export class DifyHandler implements ApiHandler {
 		}
 	}
 
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		console.log("[DIFY DEBUG] createMessage called with:", {
 			systemPromptLength: systemPrompt?.length || 0,
 			messagesCount: messages?.length || 0,
@@ -384,7 +384,7 @@ export class DifyHandler implements ApiHandler {
 		}
 	}
 
-	private convertMessagesToQuery(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): string {
+	private convertMessagesToQuery(systemPrompt: string, messages: ClineStorageMessage[]): string {
 		// Dify's context is managed by `conversation_id`. The `query` should be the last user message.
 		// The system prompt is typically configured in the Dify App itself.
 		const lastUserMessage = messages.filter((m) => m.role === "user").pop()

src/core/api/providers/doubao.ts

@@ -1,6 +1,6 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { DoubaoModelId, doubaoDefaultModelId, doubaoModels, ModelInfo } from "@shared/api"
 import OpenAI from "openai"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from ".."
 import { withRetry } from "../retry"
@@ -50,7 +50,7 @@ export class DoubaoHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [

src/core/api/providers/fireworks.ts

@@ -1,6 +1,6 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { FireworksModelId, fireworksDefaultModelId, fireworksModels, ModelInfo } from "@shared/api"
 import OpenAI from "openai"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from ".."
 import { withRetry } from "../retry"
@@ -41,7 +41,7 @@ export class FireworksHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		const client = this.ensureClient()
 		const modelId = this.options.fireworksModelId ?? ""
 

src/core/api/providers/gemini.ts

@@ -1,4 +1,3 @@
-import type { Anthropic } from "@anthropic-ai/sdk"
 // Restore GenerateContentConfig import and add GenerateContentResponseUsageMetadata
 import {
 	ApiError,
@@ -7,10 +6,11 @@ import {
 	type GenerateContentResponseUsageMetadata,
 	GoogleGenAI,
 	FunctionDeclaration as GoogleTool,
-	Part,
+	ThinkingLevel,
 } from "@google/genai"
 import { GeminiModelId, geminiDefaultModelId, geminiModels, ModelInfo } from "@shared/api"
 import { telemetryService } from "@/services/telemetry"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { RetriableError, withRetry } from "../retry"
 import { convertAnthropicMessageToGemini } from "../transform/gemini-format"
@@ -28,6 +28,7 @@ interface GeminiHandlerOptions extends CommonApiHandlerOptions {
 	geminiApiKey?: string
 	geminiBaseUrl?: string
 	thinkingBudgetTokens?: number
+	thinkingLevel?: string
 	apiModelId?: string
 	ulid?: string
 }
@@ -110,34 +111,49 @@ export class GeminiHandler implements ApiHandler {
 		baseDelay: 2000,
 		maxDelay: 15000,
 	})
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: GoogleTool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: GoogleTool[]): ApiStream {
 		const client = this.ensureClient()
 		const { id: modelId, info } = this.getModel()
 		const contents = messages.map(convertAnthropicMessageToGemini)
 
 		// Configure thinking budget if supported
-		const thinkingBudget = this.options.thinkingBudgetTokens ?? 0
-		const _maxBudget = info.thinkingConfig?.maxBudget ?? 0
+		const _thinkingBudget = this.options.thinkingBudgetTokens ?? 0
+		const maxBudget = info.thinkingConfig?.maxBudget ?? 24576
+		const thinkingBudget = Math.min(_thinkingBudget, maxBudget)
+		// When ThinkingLevel is defineded, thinking budget cannot be zero
+		// and only level is used to control thinking behavior.
+		let thinkingLevel: ThinkingLevel | undefined
+		if (this.options.thinkingLevel === "low") {
+			thinkingLevel = ThinkingLevel.LOW
+		} else if (this.options.thinkingLevel === "high") {
+			thinkingLevel = ThinkingLevel.HIGH
+		}
 
 		// Set up base generation config
 		const requestConfig: GenerateContentConfig = {
 			// Add base URL if configured
 			httpOptions: this.options.geminiBaseUrl ? { baseUrl: this.options.geminiBaseUrl } : undefined,
-			...{ systemInstruction: systemPrompt },
+			systemInstruction: systemPrompt,
 			// Set temperature (default to 0)
-			temperature: 0,
+			// Gemini 3.0 recommends 1.0
+			temperature: info.temperature ?? 1,
 		}
 
 		// Add thinking config if the model supports it
-		if (thinkingBudget > 0) {
-			requestConfig.thinkingConfig = {
-				thinkingBudget: thinkingBudget,
-				includeThoughts: true,
-			}
+		requestConfig.thinkingConfig = {
+			// Turn off thinking:
+			// thinkingBudget: 0
+			// Turn on dynamic thinking:
+			// thinkingBudget: -1
+			// Turn on fixed thinking budget:
+			thinkingBudget: thinkingLevel ? undefined : thinkingBudget,
+			thinkingLevel,
+			includeThoughts: thinkingBudget > 0 || !!thinkingLevel,
 		}
 
 		// Generate content using the configured parameters
 		const sdkCallStartTime = Date.now()
+		let responseId: string | undefined
 		let sdkFirstChunkTime: number | undefined
 		let ttftSdkMs: number | undefined
 		let apiSuccess = false
@@ -148,7 +164,8 @@ export class GeminiHandler implements ApiHandler {
 		let thoughtsTokenCount = 0 // Initialize thought token counts
 		let lastUsageMetadata: GenerateContentResponseUsageMetadata | undefined
 
-		if (tools?.length) {
+		const isNativeToolCallsEnabled = tools?.length
+		if (isNativeToolCallsEnabled) {
 			requestConfig.tools = [{ functionDeclarations: tools }]
 			requestConfig.toolConfig = {
 				// Force the model to call 'any' function.
@@ -176,56 +193,45 @@ export class GeminiHandler implements ApiHandler {
 				}
 
 				// Handle thinking content from Gemini's response
-				const candidateForThoughts = chunk?.candidates?.[0]
-				const partsForThoughts = candidateForThoughts?.content?.parts
-				let thoughts = "" // Initialize as empty string
-
-				if (partsForThoughts) {
-					// This ensures partsForThoughts is a Part[] array
-					for (const part of partsForThoughts) {
-						const { thought, text } = part as Part
-						if (thought && text) {
-							// Ensure part.text exists
-							// Handle the thought part
-							thoughts += text + "\n" // Append thought and a newline
+				const parts = chunk?.candidates?.[0]?.content?.parts || []
+				for (const part of parts) {
+					if (part.thought && part.text) {
+						yield {
+							type: "reasoning",
+							id: chunk.responseId,
+							reasoning: part.text || "",
+							signature: part.thoughtSignature,
+						}
+					} else if (part.text) {
+						yield {
+							type: "text",
+							text: part.text,
+							id: chunk.responseId,
+							signature: part.thoughtSignature,
 						}
 					}
-				}
-
-				if (thoughts.trim() !== "") {
-					yield {
-						type: "reasoning",
-						reasoning: thoughts.trim(),
-					}
-					thoughts = "" // Reset thoughts after yielding
-				}
-
-				if (chunk.text) {
-					yield {
-						type: "text",
-						text: chunk.text,
-					}
-				}
-
-				if (tools && chunk.functionCalls && chunk.functionCalls?.length > 0) {
-					for (const functionCall of chunk.functionCalls) {
-						if (functionCall.args) {
-							console.log("[GeminiHandler] tool call received:", functionCall)
+					if (part.functionCall) {
+						const functionCall = part.functionCall
+						const args = Object.entries(functionCall.args || {}).filter(([_key, val]) => !!val)
+						if (functionCall.args && args.length > 0) {
 							yield {
 								type: "tool_calls",
+								id: chunk.responseId,
 								tool_call: {
 									function: {
-										id: functionCall.id || functionCall.name,
+										id: chunk.responseId,
 										name: functionCall.name,
 										arguments: JSON.stringify(functionCall.args),
 									},
 								},
+								signature: part.thoughtSignature,
 							}
 						}
 					}
 				}
 
 				if (chunk.usageMetadata) {
+					responseId = chunk.responseId
 					lastUsageMetadata = chunk.usageMetadata
 					promptTokens = lastUsageMetadata.promptTokenCount ?? promptTokens
 					outputTokens = lastUsageMetadata.candidatesTokenCount ?? outputTokens
@@ -251,6 +257,7 @@ export class GeminiHandler implements ApiHandler {
 					cacheReadTokens,
 					cacheWriteTokens: 0,
 					totalCost,
+					id: responseId,
 				}
 			}
 		} catch (error) {

src/core/api/providers/groq.ts

@@ -1,8 +1,8 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { GroqModelId, groqDefaultModelId, groqModels, ModelInfo } from "@shared/api"
 import { calculateApiCostOpenAI } from "@utils/cost"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -192,7 +192,7 @@ export class GroqHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 		const modelFamily = this.detectModelFamily(model.id)

src/core/api/providers/hicap.ts

@@ -1,7 +1,7 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { hicapModelInfoSaneDefaults, ModelInfo } from "@shared/api"
 import OpenAI from "openai"
 import type { ChatCompletionReasoningEffort } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
 import { withRetry } from "../retry"
 import { convertToOpenAiMessages } from "../transform/openai-format"
@@ -44,7 +44,7 @@ export class HicapHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		const client = this.ensureClient()
 		const modelId = this.options.hicapModelId ?? ""
 

src/core/api/providers/huawei-cloud-maas.ts

@@ -1,7 +1,7 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { HuaweiCloudMaasModelId, huaweiCloudMaasDefaultModelId, huaweiCloudMaasModels, ModelInfo } from "@shared/api"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from ".."
 import { withRetry } from "../retry"
@@ -62,7 +62,7 @@ export class HuaweiCloudMaaSHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [

src/core/api/providers/huggingface.ts

@@ -1,8 +1,8 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { HuggingFaceModelId, huggingFaceDefaultModelId, huggingFaceModels, ModelInfo } from "@shared/api"
 import { calculateApiCostOpenAI } from "@utils/cost"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -69,7 +69,7 @@ export class HuggingFaceHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		try {
 			const client = this.ensureClient()
 			const model = this.getModel()

src/core/api/providers/litellm.ts

@@ -1,6 +1,7 @@
 import { Anthropic } from "@anthropic-ai/sdk"
 import { LiteLLMModelInfo, liteLlmDefaultModelId, liteLlmModelInfoSaneDefaults } from "@shared/api"
 import OpenAI from "openai"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { isAnthropicModelId } from "@/utils/model-utils"
 import { ApiHandler, CommonApiHandlerOptions } from ".."
@@ -183,7 +184,7 @@ export class LiteLlmHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		const client = this.ensureClient()
 		const formattedMessages = convertToOpenAiMessages(messages)
 		const systemMessage: OpenAI.Chat.ChatCompletionSystemMessageParam | Anthropic.Messages.TextBlockParam = {

src/core/api/providers/lmstudio.ts

@@ -1,7 +1,7 @@
-import type { Anthropic } from "@anthropic-ai/sdk"
 import { type ModelInfo, openAiModelInfoSaneDefaults } from "@shared/api"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import type { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -40,7 +40,7 @@ export class LmStudioHandler implements ApiHandler {
 	}
 
 	@withRetry({ retryAllErrors: true })
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [
 			{ role: "system", content: systemPrompt },

src/core/api/providers/minimax.ts

@@ -2,6 +2,7 @@ import { Anthropic } from "@anthropic-ai/sdk"
 import { Tool as AnthropicTool } from "@anthropic-ai/sdk/resources/index"
 import { Stream as AnthropicStream } from "@anthropic-ai/sdk/streaming"
 import { MinimaxModelId, ModelInfo, minimaxDefaultModelId, minimaxModels } from "@/shared/api"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ClineTool } from "@/shared/tools"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
@@ -45,7 +46,7 @@ export class MinimaxHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: ClineTool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: ClineTool[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 
@@ -66,12 +67,11 @@ export class MinimaxHandler implements ApiHandler {
 			tool_choice: nativeToolsOn ? { type: "any" } : undefined,
 		})
 
-		let thinkingDeltaAccumulator = ""
 		const lastStartedToolCall = { id: "", name: "", arguments: "" }
 
 		for await (const chunk of stream) {
 			switch (chunk?.type) {
-				case "message_start":
+				case "message_start": {
 					// tells us cache reads/writes/input/output
 					const usage = chunk.message.usage
 					yield {
@@ -82,6 +82,7 @@ export class MinimaxHandler implements ApiHandler {
 						cacheReadTokens: usage.cache_read_input_tokens || undefined,
 					}
 					break
+				}
 				case "message_delta":
 					// tells us stop_reason, stop_sequence, and output tokens along the way and at the end of the message
 					yield {
@@ -100,13 +101,11 @@ export class MinimaxHandler implements ApiHandler {
 								type: "reasoning",
 								reasoning: chunk.content_block.thinking || "",
 							}
-							const thinking = chunk.content_block.thinking
-							const signature = chunk.content_block.signature
-							if (thinking && signature) {
+							if (chunk.content_block.thinking && chunk.content_block.signature) {
 								yield {
-									type: "ant_thinking",
-									thinking,
-									signature,
+									type: "reasoning",
+									reasoning: chunk.content_block.thinking,
+									signature: chunk.content_block.signature,
 								}
 							}
 							break
@@ -115,10 +114,7 @@ export class MinimaxHandler implements ApiHandler {
 							yield {
 								type: "reasoning",
 								reasoning: "[Redacted thinking block]",
-							}
-							yield {
-								type: "ant_redacted_thinking",
-								data: chunk.content_block.data,
+								redacted_data: chunk.content_block.data,
 							}
 							break
 						case "tool_use":
@@ -147,20 +143,19 @@ export class MinimaxHandler implements ApiHandler {
 				case "content_block_delta":
 					switch (chunk.delta.type) {
 						case "thinking_delta":
-							// 'reasoning' type just displays in the UI, but ant_thinking will be used to send the thinking traces back to the API
+							// 'reasoning' type just displays in the UI, but reasoning with signature will be used to send the thinking traces back to the API
 							yield {
 								type: "reasoning",
 								reasoning: chunk.delta.thinking,
 							}
-							thinkingDeltaAccumulator += chunk.delta.thinking
 							break
 						case "signature_delta":
 							// It's used when sending the thinking block back to the API
 							// API expects this in completed form, not as array of deltas
-							if (thinkingDeltaAccumulator && chunk.delta.signature) {
+							if (chunk.delta.signature) {
 								yield {
-									type: "ant_thinking",
-									thinking: thinkingDeltaAccumulator,
+									type: "reasoning",
+									reasoning: "",
 									signature: chunk.delta.signature,
 								}
 							}

src/core/api/providers/mistral.ts

@@ -1,9 +1,9 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { Mistral } from "@mistralai/mistralai"
 import { HTTPClient } from "@mistralai/mistralai/lib/http"
 import { Tool as MistralTool } from "@mistralai/mistralai/models/components/tool"
 import { MistralModelId, ModelInfo, mistralDefaultModelId, mistralModels } from "@shared/api"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -48,7 +48,7 @@ export class MistralHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const stream = await client.chat
 			.stream({

src/core/api/providers/moonshot.ts

@@ -1,7 +1,7 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
 import { ModelInfo, MoonshotModelId, moonshotDefaultModelId, moonshotModels } from "@/shared/api"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
 import { withRetry } from "../retry"
@@ -40,7 +40,7 @@ export class MoonshotHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 

src/core/api/providers/nebius.ts

@@ -1,7 +1,7 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { type ModelInfo, type NebiusModelId, nebiusDefaultModelId, nebiusModels } from "@shared/api"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
 import { withRetry } from "../retry"
@@ -39,7 +39,7 @@ export class NebiusHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 

src/core/api/providers/nousresearch.ts

@@ -0,0 +1,92 @@
+import { ModelInfo, NousResearchModelId, nousResearchDefaultModelId, nousResearchModels } from "@shared/api"
+import OpenAI from "openai"
+import { ClineStorageMessage } from "@/shared/messages/content"
+import { ApiHandler, CommonApiHandlerOptions } from "../index"
+import { withRetry } from "../retry"
+import { convertToOpenAiMessages } from "../transform/openai-format"
+import { ApiStream } from "../transform/stream"
+
+interface NousResearchHandlerOptions extends CommonApiHandlerOptions {
+	nousResearchApiKey?: string
+	apiModelId?: string
+}
+
+export class NousResearchHandler implements ApiHandler {
+	private options: NousResearchHandlerOptions
+	private client: OpenAI | undefined
+
+	constructor(options: NousResearchHandlerOptions) {
+		this.options = options
+	}
+
+	private ensureClient(): OpenAI {
+		if (!this.client) {
+			if (!this.options.nousResearchApiKey) {
+				throw new Error("NousResearch API key is required")
+			}
+			try {
+				this.client = new OpenAI({
+					baseURL: "https://inference-api.nousResearch.com/v1",
+					apiKey: this.options.nousResearchApiKey,
+				})
+			} catch (error: any) {
+				throw new Error(`Error creating NousResearch client: ${error.message}`)
+			}
+		}
+		return this.client
+	}
+
+	@withRetry()
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
+		const client = this.ensureClient()
+		const model = this.getModel()
+
+		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [
+			{ role: "system", content: systemPrompt },
+			...convertToOpenAiMessages(messages),
+		]
+
+		const stream = await client.chat.completions.create({
+			model: model.id,
+			messages: openAiMessages,
+			temperature: 0,
+			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,
+				}
+			}
+
+			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,
+				}
+			}
+		}
+	}
+
+	getModel(): { id: NousResearchModelId; info: ModelInfo } {
+		const modelId = this.options.apiModelId
+
+		if (modelId && modelId in nousResearchModels) {
+			const id = modelId as NousResearchModelId
+			return { id, info: nousResearchModels[id] }
+		}
+		return { id: nousResearchDefaultModelId, info: nousResearchModels[nousResearchDefaultModelId] }
+	}
+}

src/core/api/providers/oca.ts

@@ -1,7 +1,5 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { LiteLLMModelInfo, liteLlmDefaultModelId, liteLlmModelInfoSaneDefaults } from "@shared/api"
 import OpenAI, { APIError, OpenAIError } from "openai"
-import type { FinalRequestOptions, Headers as OpenAIHeaders } from "openai/core"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
 import { OcaAuthService } from "@/services/auth/oca/OcaAuthService"
 import {
@@ -11,6 +9,7 @@ import {
 } from "@/services/auth/oca/utils/constants"
 import { createOcaHeaders } from "@/services/auth/oca/utils/utils"
 import { Logger } from "@/services/logging/Logger"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, type CommonApiHandlerOptions } from ".."
 import { withRetry } from "../retry"
@@ -38,7 +37,7 @@ export class OcaHandler implements ApiHandler {
 
 	protected initializeClient(options: OcaHandlerOptions) {
 		return new (class OCIOpenAI extends OpenAI {
-			protected override async prepareOptions(opts: FinalRequestOptions<unknown>): Promise<void> {
+			protected override async prepareOptions(opts: any): Promise<void> {
 				const token = await OcaAuthService.getInstance().getAuthToken()
 				if (!token) {
 					throw new OpenAIError("Unable to handle auth, Oracle Code Assist (OCA) access token is not available")
@@ -55,7 +54,7 @@ export class OcaHandler implements ApiHandler {
 				status: number | undefined,
 				error: Object | undefined,
 				message: string | undefined,
-				headers: OpenAIHeaders | undefined,
+				headers: any | undefined,
 			): APIError {
 				interface OciError {
 					code?: string
@@ -75,7 +74,8 @@ export class OcaHandler implements ApiHandler {
 				if (opcRequestId) {
 					ociErrorMessage += `\n(${OCI_HEADER_OPC_REQUEST_ID}: ${opcRequestId})`
 				}
-				return super.makeStatusError(status, error, ociErrorMessage, headers)
+				const statusCode = typeof status === "number" ? status : 500
+				return super.makeStatusError(statusCode, error ?? {}, ociErrorMessage, headers)
 			}
 		})({
 			baseURL:
@@ -139,7 +139,7 @@ export class OcaHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const formattedMessages = convertToOpenAiMessages(messages)
 		const systemMessage: OpenAI.Chat.ChatCompletionSystemMessageParam = {
@@ -151,7 +151,7 @@ export class OcaHandler implements ApiHandler {
 
 		// Configuration for extended thinking
 		const budgetTokens = this.options.thinkingBudgetTokens || 0
-		const reasoningOn = budgetTokens !== 0 ? true : false
+		const reasoningOn = budgetTokens !== 0
 		const thinkingConfig = reasoningOn ? { type: "enabled", budget_tokens: budgetTokens } : undefined
 
 		let temperature: number | undefined = this.options.ocaModelInfo?.temperature ?? 0

src/core/api/providers/ollama.ts

@@ -1,6 +1,6 @@
-import type { Anthropic } from "@anthropic-ai/sdk"
 import { type ModelInfo, openAiModelInfoSaneDefaults } from "@shared/api"
 import { type Config, type Message, Ollama } from "ollama"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import type { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
 import { convertToOllamaMessages } from "../transform/ollama-format"
@@ -48,7 +48,7 @@ export class OllamaHandler implements ApiHandler {
 	}
 
 	@withRetry({ retryAllErrors: true })
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		const client = this.ensureClient()
 		const ollamaMessages: Message[] = [{ role: "system", content: systemPrompt }, ...convertToOllamaMessages(messages)]
 

src/core/api/providers/openai-native.ts

@@ -1,12 +1,14 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ModelInfo, OpenAiNativeModelId, openAiNativeDefaultModelId, openAiNativeModels } from "@shared/api"
 import { calculateApiCostOpenAI } from "@utils/cost"
 import OpenAI from "openai"
 import type { ChatCompletionReasoningEffort, ChatCompletionTool } from "openai/resources/chat/completions"
+import { Logger } from "@/services/logging/Logger"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
 import { convertToOpenAiMessages } from "../transform/openai-format"
+import { convertToOpenAIResponsesInput } from "../transform/openai-response-format"
 import { ApiStream } from "../transform/stream"
 import { getOpenAIToolParams, ToolCallProcessor } from "../transform/tool-call-processor"
 
@@ -61,7 +63,20 @@ export class OpenAiNativeHandler implements ApiHandler {
 	@withRetry()
 	async *createMessage(
 		systemPrompt: string,
-		messages: Anthropic.Messages.MessageParam[],
+		messages: ClineStorageMessage[],
+		tools?: ChatCompletionTool[],
+		useResponseFormat = false,
+	): ApiStream {
+		if (useResponseFormat) {
+			yield* this.createResponseStream(systemPrompt, messages, tools)
+		} else {
+			yield* this.createCompletionStream(systemPrompt, messages, tools)
+		}
+	}
+
+	private async *createCompletionStream(
+		systemPrompt: string,
+		messages: ClineStorageMessage[],
 		tools?: ChatCompletionTool[],
 	): ApiStream {
 		const client = this.ensureClient()
@@ -115,6 +130,9 @@ export class OpenAiNativeHandler implements ApiHandler {
 			case "gpt-5-2025-08-07":
 			case "gpt-5-mini-2025-08-07":
 			case "gpt-5-nano-2025-08-07":
+			case "gpt-5.1-2025-11-13":
+			case "gpt-5.1-chat-latest":
+			case "gpt-5.1": {
 				const stream = await client.chat.completions.create({
 					model: model.id,
 					temperature: 1,
@@ -148,6 +166,7 @@ export class OpenAiNativeHandler implements ApiHandler {
 					}
 				}
 				break
+			}
 			default: {
 				const stream = await client.chat.completions.create({
 					model: model.id,
@@ -181,6 +200,208 @@ export class OpenAiNativeHandler implements ApiHandler {
 		}
 	}
 
+	private async *createResponseStream(
+		systemPrompt: string,
+		messages: ClineStorageMessage[],
+		tools?: ChatCompletionTool[],
+	): ApiStream {
+		const client = this.ensureClient()
+		const model = this.getModel()
+
+		// Convert messages to Responses API input format
+		const input = convertToOpenAIResponsesInput(messages)
+
+		// Convert ChatCompletion tools to Responses API format if provided
+		const responseTools = tools
+			?.filter((tool) => tool.type === "function")
+			.map((tool: any) => ({
+				type: "function" as const,
+				name: tool.function.name,
+				description: tool.function.description,
+				parameters: tool.function.parameters,
+				strict: tool.function.strict ?? true, // Responses API defaults to strict mode
+			}))
+
+		Logger.debug("OpenAI Responses Input: " + JSON.stringify(input))
+
+		// const lastAssistantMessage = [...messages].reverse().find((msg) => msg.role === "assistant" && msg.id)
+		// const previous_response_id = lastAssistantMessage?.id
+
+		// Create the response using Responses API
+		const stream = await client.responses.create({
+			model: model.id,
+			instructions: systemPrompt,
+			input,
+			stream: true,
+			tools: responseTools,
+			// previous_response_id,
+			// store: true,
+			reasoning: { effort: "medium", summary: "auto" },
+			// include: ["reasoning.encrypted_content"],
+		})
+
+		// Process the response stream
+		for await (const chunk of stream) {
+			Logger.debug("OpenAI Responses Chunk: " + JSON.stringify(chunk))
+
+			// Handle different event types from Responses API
+			if (chunk.type === "response.output_item.added") {
+				const item = chunk.item
+				if (item.type === "function_call" && item.id) {
+					yield {
+						type: "tool_calls",
+						id: item.id,
+						tool_call: {
+							call_id: item.call_id,
+							function: {
+								id: item.id,
+								name: item.name,
+								arguments: item.arguments,
+							},
+						},
+					}
+				}
+				if (item.type === "reasoning" && item.encrypted_content && item.id) {
+					yield {
+						type: "reasoning",
+						id: item.id,
+						reasoning: "",
+						redacted_data: item.encrypted_content,
+					}
+				}
+			}
+			if (chunk.type === "response.output_item.done") {
+				const item = chunk.item
+				if (item.type === "function_call") {
+					yield {
+						type: "tool_calls",
+						id: item.id || item.call_id,
+						tool_call: {
+							call_id: item.call_id,
+							function: {
+								id: item.id,
+								name: item.name,
+								arguments: item.arguments,
+							},
+						},
+					}
+				}
+				if (item.type === "reasoning") {
+					yield {
+						type: "reasoning",
+						id: item.id,
+						details: item.summary,
+						reasoning: "",
+					}
+				}
+			}
+			if (chunk.type === "response.reasoning_summary_part.added") {
+				yield {
+					type: "reasoning",
+					id: chunk.item_id,
+					reasoning: chunk.part.text,
+				}
+			}
+			if (chunk.type === "response.reasoning_summary_text.delta") {
+				yield {
+					type: "reasoning",
+					id: chunk.item_id,
+					reasoning: chunk.delta,
+				}
+			}
+			if (chunk.type === "response.reasoning_summary_part.done") {
+				yield {
+					type: "reasoning",
+					id: chunk.item_id,
+					details: chunk.part,
+					reasoning: "",
+				}
+			}
+			if (chunk.type === "response.output_text.delta") {
+				// Handle text content deltas
+				if (chunk.delta) {
+					yield {
+						id: chunk.item_id,
+						type: "text",
+						text: chunk.delta,
+					}
+				}
+			}
+			if (chunk.type === "response.reasoning_text.delta") {
+				// Handle reasoning content deltas
+				if (chunk.delta) {
+					yield {
+						id: chunk.item_id,
+						type: "reasoning",
+						reasoning: chunk.delta,
+					}
+				}
+			}
+			if (chunk.type === "response.function_call_arguments.delta") {
+				yield {
+					type: "tool_calls",
+					tool_call: {
+						function: {
+							id: chunk.item_id,
+							name: chunk.item_id,
+							arguments: chunk.delta,
+						},
+					},
+				}
+			}
+			if (chunk.type === "response.function_call_arguments.done") {
+				// Handle completed function call
+				if (chunk.item_id && chunk.name && chunk.arguments) {
+					yield {
+						type: "tool_calls",
+						tool_call: {
+							function: {
+								id: chunk.item_id,
+								name: chunk.name,
+								arguments: chunk.arguments,
+							},
+						},
+					}
+				}
+			}
+
+			if (
+				chunk.type === "response.incomplete" &&
+				chunk.response?.status === "incomplete" &&
+				chunk.response?.incomplete_details?.reason === "max_output_tokens"
+			) {
+				console.log("Ran out of tokens")
+				if (chunk.response?.output_text?.length > 0) {
+					console.log("Partial output:", chunk.response.output_text)
+				} else {
+					console.log("Ran out of tokens during reasoning")
+				}
+			}
+
+			if (chunk.type === "response.completed" && chunk.response?.usage) {
+				// Handle usage information when response is complete
+				const usage = chunk.response.usage
+				const inputTokens = usage.input_tokens || 0
+				const outputTokens = usage.output_tokens || 0
+				const cacheReadTokens = usage.output_tokens_details?.reasoning_tokens || 0
+				const cacheWriteTokens = usage.input_tokens_details?.cached_tokens || 0
+				const totalTokens = usage.total_tokens || 0
+				Logger.log(`Total tokens from Responses API usage: ${totalTokens}`)
+				const totalCost = calculateApiCostOpenAI(model.info, inputTokens, outputTokens, cacheWriteTokens, cacheReadTokens)
+				const nonCachedInputTokens = Math.max(0, inputTokens - cacheReadTokens - cacheWriteTokens)
+				yield {
+					type: "usage",
+					inputTokens: nonCachedInputTokens,
+					outputTokens: outputTokens,
+					cacheWriteTokens: cacheWriteTokens,
+					cacheReadTokens: cacheReadTokens,
+					totalCost: totalCost,
+					id: chunk.response.id,
+				}
+			}
+		}
+	}
+
 	getModel(): { id: OpenAiNativeModelId; info: ModelInfo } {
 		const modelId = this.options.apiModelId
 		if (modelId && modelId in openAiNativeModels) {

src/core/api/providers/openai.ts

@@ -1,7 +1,7 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { azureOpenAiDefaultApiVersion, ModelInfo, OpenAiCompatibleModelInfo, openAiModelInfoSaneDefaults } from "@shared/api"
 import OpenAI, { AzureOpenAI } from "openai"
 import type { ChatCompletionReasoningEffort, ChatCompletionTool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
 import { withRetry } from "../retry"
@@ -65,11 +65,7 @@ export class OpenAiHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(
-		systemPrompt: string,
-		messages: Anthropic.Messages.MessageParam[],
-		tools?: ChatCompletionTool[],
-	): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: ChatCompletionTool[]): ApiStream {
 		const client = this.ensureClient()
 		const modelId = this.options.openAiModelId ?? ""
 		const isDeepseekReasoner = modelId.includes("deepseek-reasoner")
@@ -81,7 +77,13 @@ export class OpenAiHandler implements ApiHandler {
 			{ role: "system", content: systemPrompt },
 			...convertToOpenAiMessages(messages),
 		]
-		let temperature: number | undefined = this.options.openAiModelInfo?.temperature ?? openAiModelInfoSaneDefaults.temperature
+		let temperature: number | undefined
+		if (this.options.openAiModelInfo?.temperature !== undefined) {
+			const tempValue = Number(this.options.openAiModelInfo.temperature)
+			temperature = tempValue === 0 ? undefined : tempValue
+		} else {
+			temperature = openAiModelInfoSaneDefaults.temperature
+		}
 		let reasoningEffort: ChatCompletionReasoningEffort | undefined
 		let maxTokens: number | undefined
 

src/core/api/providers/openrouter.ts

@@ -1,10 +1,11 @@
 import { setTimeout as setTimeoutPromise } from "node:timers/promises"
-import { Anthropic } from "@anthropic-ai/sdk"
+import { StateManager } from "@core/storage/StateManager"
 import { ModelInfo, openRouterDefaultModelId, openRouterDefaultModelInfo } from "@shared/api"
 import { shouldSkipReasoningForModel } from "@utils/model-utils"
 import axios from "axios"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch, getAxiosSettings } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -20,6 +21,7 @@ interface OpenRouterHandlerOptions extends CommonApiHandlerOptions {
 	openRouterProviderSorting?: string
 	reasoningEffort?: string
 	thinkingBudgetTokens?: number
+	geminiThinkingLevel?: string
 }
 
 export class OpenRouterHandler implements ApiHandler {
@@ -54,7 +56,7 @@ export class OpenRouterHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		this.lastGenerationId = undefined
 
@@ -67,6 +69,7 @@ export class OpenRouterHandler implements ApiHandler {
 			this.options.thinkingBudgetTokens,
 			this.options.openRouterProviderSorting,
 			tools,
+			this.options.geminiThinkingLevel,
 		)
 
 		let didOutputUsage: boolean = false
@@ -127,8 +130,7 @@ export class OpenRouterHandler implements ApiHandler {
 			if ("reasoning" in delta && delta.reasoning && !shouldSkipReasoningForModel(this.options.openRouterModelId)) {
 				yield {
 					type: "reasoning",
-					// @ts-ignore-next-line
-					reasoning: delta.reasoning,
+					reasoning: typeof delta.reasoning === "string" ? delta.reasoning : JSON.stringify(delta.reasoning),
 				}
 			}
 
@@ -142,8 +144,9 @@ export class OpenRouterHandler implements ApiHandler {
 				!shouldSkipReasoningForModel(this.options.openRouterModelId)
 			) {
 				yield {
-					type: "reasoning_details",
-					reasoning_details: delta.reasoning_details,
+					type: "reasoning",
+					reasoning: "",
+					details: delta.reasoning_details,
 				}
 			}
 
@@ -214,11 +217,11 @@ export class OpenRouterHandler implements ApiHandler {
 	}
 
 	getModel(): { id: string; info: ModelInfo } {
-		const modelId = this.options.openRouterModelId
-		const modelInfo = this.options.openRouterModelInfo
-		if (modelId && modelInfo) {
-			return { id: modelId, info: modelInfo }
+		const modelId = this.options.openRouterModelId || openRouterDefaultModelId
+		const cachedModelInfo = StateManager.get().getModelInfo("openRouter", modelId)
+		return {
+			id: modelId,
+			info: cachedModelInfo || openRouterDefaultModelInfo,
 		}
-		return { id: openRouterDefaultModelId, info: openRouterDefaultModelInfo }
 	}
 }

src/core/api/providers/qwen-code.ts

@@ -1,10 +1,10 @@
 import { promises as fs } from "node:fs"
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ModelInfo, QwenCodeModelId, qwenCodeDefaultModelId, qwenCodeModels } from "@shared/api"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
 import * as os from "os"
 import * as path from "path"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -177,7 +177,7 @@ export class QwenCodeHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		await this.ensureAuthenticated()
 		const client = this.ensureClient()
 		const model = this.getModel()

src/core/api/providers/qwen.ts

@@ -1,4 +1,3 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import {
 	InternationalQwenModelId,
 	internationalQwenDefaultModelId,
@@ -11,6 +10,7 @@ import {
 } from "@shared/api"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -81,7 +81,7 @@ export class QwenHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 		const isDeepseekReasoner = model.id.includes("deepseek-r1")

src/core/api/providers/requesty.ts

@@ -1,8 +1,8 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ModelInfo, requestyDefaultModelId, requestyDefaultModelInfo } from "@shared/api"
 import { calculateApiCostOpenAI } from "@utils/cost"
 import OpenAI from "openai"
 import { toRequestyServiceStringUrl } from "@/shared/clients/requesty"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
 import { withRetry } from "../retry"
@@ -59,7 +59,7 @@ export class RequestyHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 

src/core/api/providers/sambanova.ts

@@ -1,7 +1,7 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ModelInfo, SambanovaModelId, sambanovaDefaultModelId, sambanovaModels } from "@shared/api"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
 import { withRetry } from "../retry"
@@ -42,7 +42,7 @@ export class SambanovaHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 

src/core/api/providers/sapaicore.ts

@@ -4,10 +4,11 @@ import {
 	ConversationRole as BedrockConversationRole,
 	type Message as BedrockMessage,
 } from "@aws-sdk/client-bedrock-runtime"
-import { ChatMessages, LlmModuleConfig, OrchestrationClient, TemplatingModuleConfig } from "@sap-ai-sdk/orchestration"
+import { ChatMessage, OrchestrationClient, OrchestrationModuleConfig } from "@sap-ai-sdk/orchestration"
 import { ModelInfo, SapAiCoreModelId, sapAiCoreDefaultModelId, sapAiCoreModels } from "@shared/api"
 import axios from "axios"
 import OpenAI from "openai"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { getAxiosSettings } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -115,7 +116,7 @@ namespace Bedrock {
 	 * Formats messages for models using the Converse API specification
 	 * Used by both Anthropic and Nova models to avoid code duplication
 	 */
-	export function formatMessagesForConverseAPI(messages: Anthropic.Messages.MessageParam[]): BedrockMessage[] {
+	export function formatMessagesForConverseAPI(messages: ClineStorageMessage[]): BedrockMessage[] {
 		return messages.map((message) => {
 			// Determine role (user or assistant)
 			const role = message.role === "user" ? BedrockConversationRole.USER : BedrockConversationRole.ASSISTANT
@@ -315,7 +316,7 @@ namespace Gemini {
 	 */
 	export function prepareRequestPayload(
 		systemPrompt: string,
-		messages: Anthropic.Messages.MessageParam[],
+		messages: ClineStorageMessage[],
 		model: { id: SapAiCoreModelId; info: ModelInfo },
 		thinkingBudgetTokens?: number,
 	): any {
@@ -458,7 +459,7 @@ export class SapAiCoreHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		if (this.options.sapAiCoreUseOrchestrationMode) {
 			yield* this.createMessageWithOrchestration(systemPrompt, messages)
 		} else {
@@ -490,29 +491,31 @@ export class SapAiCoreHandler implements ApiHandler {
 		this.isAiCoreEnvSetup = true
 	}
 
-	private async *createMessageWithOrchestration(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	private async *createMessageWithOrchestration(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		try {
 			// Ensure AI Core environment variable is set up (only runs once)
 			this.ensureAiCoreEnvSetup()
 			const model = this.getModel()
 
-			// Define the LLM to be used by the Orchestration pipeline
-			const llm: LlmModuleConfig = {
-				model_name: model.id,
+			const orchestrationConfig: OrchestrationModuleConfig = {
+				promptTemplating: {
+					model: {
+						name: model.id,
+					},
+					prompt: {
+						template: [
+							{
+								role: "system",
+								content: systemPrompt,
+							},
+						],
+					},
+				},
 			}
 
-			const templating: TemplatingModuleConfig = {
-				template: [
-					{
-						role: "system",
-						content: systemPrompt,
-					},
-				],
-			}
-			const orchestrationClient = new OrchestrationClient(
-				{ llm, templating },
-				{ resourceGroup: this.options.sapAiResourceGroup || "default" },
-			)
+			const orchestrationClient = new OrchestrationClient(orchestrationConfig, {
+				resourceGroup: this.options.sapAiResourceGroup || "default",
+			})
 
 			const sapMessages = this.convertMessageParamToSAPMessages(messages)
 
@@ -538,7 +541,7 @@ export class SapAiCoreHandler implements ApiHandler {
 		}
 	}
 
-	private async *createMessageWithDeployments(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	private async *createMessageWithDeployments(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		const token = await this.getToken()
 		const headers = {
 			Authorization: `Bearer ${token}`,
@@ -1040,8 +1043,8 @@ export class SapAiCoreHandler implements ApiHandler {
 		}
 		return { id: sapAiCoreDefaultModelId, info: sapAiCoreModels[sapAiCoreDefaultModelId] }
 	}
-	private convertMessageParamToSAPMessages(messages: Anthropic.Messages.MessageParam[]): ChatMessages {
+	private convertMessageParamToSAPMessages(messages: ClineStorageMessage[]): ChatMessage[] {
 		// Use the existing OpenAI converter since the logic is identical
-		return convertToOpenAiMessages(messages) as ChatMessages
+		return convertToOpenAiMessages(messages) as ChatMessage[]
 	}
 }

src/core/api/providers/together.ts

@@ -1,7 +1,7 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ModelInfo, openAiModelInfoSaneDefaults } from "@shared/api"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
 import { withRetry } from "../retry"
@@ -42,7 +42,7 @@ export class TogetherHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const modelId = this.options.togetherModelId ?? ""
 		const isDeepseekReasoner = modelId.includes("deepseek-reasoner")

src/core/api/providers/vercel-ai-gateway.ts

@@ -1,7 +1,7 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ModelInfo, openRouterDefaultModelId, openRouterDefaultModelInfo } from "@shared/api"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../index"
 import { withRetry } from "../retry"
@@ -47,7 +47,7 @@ export class VercelAIGatewayHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const modelId = this.getModel().id
 		const modelInfo = this.getModel().info
@@ -82,8 +82,7 @@ export class VercelAIGatewayHandler implements ApiHandler {
 				if ("reasoning" in delta && delta.reasoning) {
 					yield {
 						type: "reasoning",
-						// @ts-ignore-next-line
-						reasoning: delta.reasoning,
+						reasoning: typeof delta.reasoning === "string" ? delta.reasoning : JSON.stringify(delta.reasoning),
 					}
 				}
 
@@ -95,28 +94,23 @@ export class VercelAIGatewayHandler implements ApiHandler {
 					delta.reasoning_details.length // exists and non-0
 				) {
 					yield {
-						type: "reasoning_details",
-						reasoning_details: delta.reasoning_details,
+						type: "reasoning",
+						reasoning: "",
+						details: delta.reasoning_details,
 					}
 				}
 
 				if (!didOutputUsage && chunk.usage) {
-					const inputTokens = chunk.usage.prompt_tokens || 0
-					const outputTokens =
-						(chunk.usage.completion_tokens || 0) + (chunk.usage.completion_tokens_details?.reasoning_tokens || 0)
-
-					const cacheReadTokens = chunk.usage.prompt_tokens_details?.cached_tokens || 0
 					// @ts-ignore - Vercel AI Gateway extends OpenAI types
-					const cacheWriteTokens = chunk.usage.cache_creation_input_tokens || 0
+					const totalCost = (chunk.usage.cost || 0) + (chunk.usage.cost_details?.upstream_inference_cost || 0)
 
 					yield {
 						type: "usage",
-						inputTokens: inputTokens,
-						outputTokens: outputTokens,
-						cacheWriteTokens: cacheWriteTokens,
-						cacheReadTokens: cacheReadTokens,
-						// @ts-expect-error - Vercel AI Gateway extends OpenAI types
-						totalCost: chunk.usage.cost || 0,
+						cacheWriteTokens: 0,
+						cacheReadTokens: chunk.usage.prompt_tokens_details?.cached_tokens || 0,
+						inputTokens: (chunk.usage.prompt_tokens || 0) - (chunk.usage.prompt_tokens_details?.cached_tokens || 0),
+						outputTokens: chunk.usage.completion_tokens || 0,
+						totalCost,
 					}
 					didOutputUsage = true
 				}

src/core/api/providers/vertex.ts

@@ -1,8 +1,8 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { Tool as AnthropicTool } from "@anthropic-ai/sdk/resources/index"
 import { AnthropicVertex } from "@anthropic-ai/vertex-sdk"
 import { FunctionDeclaration as GoogleTool } from "@google/genai"
 import { ModelInfo, VertexModelId, vertexDefaultModelId, vertexModels } from "@shared/api"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { ClineTool } from "@/shared/tools"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -67,7 +67,7 @@ export class VertexHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: ClineTool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: ClineTool[]): ApiStream {
 		const model = this.getModel()
 		const modelId = model.id
 
@@ -103,13 +103,6 @@ export class VertexHandler implements ApiHandler {
 			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 clientAnthropic.beta.messages.create(
 					{
 						model: modelId,
@@ -123,7 +116,7 @@ export class VertexHandler implements ApiHandler {
 								cache_control: { type: "ephemeral" },
 							},
 						],
-						messages: sanitizeAnthropicMessages(messages, lastUserMsgIndex, secondLastMsgUserIndex),
+						messages: sanitizeAnthropicMessages(messages, true),
 						stream: true,
 						tools: tools?.length ? (tools as AnthropicTool[]) : undefined,
 						// tool_choice options:
@@ -149,7 +142,7 @@ export class VertexHandler implements ApiHandler {
 							type: "text",
 						},
 					],
-					messages: sanitizeAnthropicMessages(messages),
+					messages: sanitizeAnthropicMessages(messages, false),
 					stream: true,
 					tools: tools?.length ? (tools as AnthropicTool[]) : undefined,
 					// tool_choice options:
@@ -166,7 +159,7 @@ export class VertexHandler implements ApiHandler {
 
 		for await (const chunk of stream) {
 			switch (chunk?.type) {
-				case "message_start":
+				case "message_start": {
 					const usage = chunk.message.usage
 					yield {
 						type: "usage",
@@ -176,6 +169,7 @@ export class VertexHandler implements ApiHandler {
 						cacheReadTokens: usage.cache_read_input_tokens || undefined,
 					}
 					break
+				}
 				case "message_delta":
 					yield {
 						type: "usage",

src/core/api/providers/vscode-lm.ts

@@ -1,8 +1,8 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ModelInfo, openAiModelInfoSaneDefaults } from "@shared/api"
 import { SELECTOR_SEPARATOR, stringifyVsCodeLmModelSelector } from "@shared/vsCodeSelectorUtils"
 import { calculateApiCostAnthropic } from "@utils/cost"
 import * as vscode from "vscode"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { ApiHandler, CommonApiHandlerOptions, SingleCompletionHandler } from "../"
 import { withRetry } from "../retry"
 import { ApiStream } from "../transform/stream"
@@ -366,7 +366,7 @@ export class VsCodeLmHandler implements ApiHandler, SingleCompletionHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[]): ApiStream {
 		// Ensure clean state before starting a new request
 		this.ensureCleanState()
 		const client: vscode.LanguageModelChat = await this.getClient()

src/core/api/providers/xai.ts

@@ -1,9 +1,9 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { ModelInfo, XAIModelId, xaiDefaultModelId, xaiModels } from "@shared/api"
 import { shouldSkipReasoningForModel } from "@utils/model-utils"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
 import { ChatCompletionReasoningEffort } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { ApiHandler, CommonApiHandlerOptions } from "../"
 import { withRetry } from "../retry"
@@ -44,7 +44,7 @@ export class XAIHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const modelId = this.getModel().id
 		// ensure reasoning effort is either "low" or "high" for grok-3-mini

src/core/api/providers/zai.ts

@@ -1,4 +1,3 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import {
 	internationalZAiDefaultModelId,
 	internationalZAiModelId,
@@ -10,6 +9,7 @@ import {
 } from "@shared/api"
 import OpenAI from "openai"
 import type { ChatCompletionTool as OpenAITool } from "openai/resources/chat/completions"
+import { ClineStorageMessage } from "@/shared/messages/content"
 import { fetch } from "@/shared/net"
 import { version as extensionVersion } from "../../../../package.json"
 import { ApiHandler, CommonApiHandlerOptions } from ".."
@@ -76,7 +76,7 @@ export class ZAiHandler implements ApiHandler {
 	}
 
 	@withRetry()
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[], tools?: OpenAITool[]): ApiStream {
+	async *createMessage(systemPrompt: string, messages: ClineStorageMessage[], tools?: OpenAITool[]): ApiStream {
 		const client = this.ensureClient()
 		const model = this.getModel()
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [

src/core/api/transform/anthropic-format.ts

@@ -1,77 +1,90 @@
-import { MessageParam } from "@anthropic-ai/sdk/resources/index"
+import { Anthropic } from "@anthropic-ai/sdk"
+import { ClineStorageMessage, convertClineStorageToAnthropicMessage } from "@/shared/messages/content"
 
 /**
- * Sanitize Anthropic messages by removing reasoning details and adding ephemeral cache control
- * to the last two user messages to prevent them from being stored in Anthropic's cache.
+ * Converts Cline storage messages to Anthropic API format with optional cache control.
+ * Adds ephemeral cache control to the last two user messages to prevent them from being
+ * stored in Anthropic's cache.
+ *
+ * @param clineMessages - Array of Cline storage messages to convert
+ * @param lastUserMsgIndex - Optional index of the last user message
+ * @param secondLastMsgUserIndex - Optional index of the second-to-last user message
+ * @returns Array of Anthropic-compatible messages with cache control applied
  */
 export function sanitizeAnthropicMessages(
-	messages: Array<MessageParam>,
-	lastUserMsgIndex?: number,
-	secondLastMsgUserIndex?: number,
-): Array<MessageParam> {
-	return messages.map((_message, index) => {
-		const message = removeReasoningDetails(_message)
-		const addCacheControl = lastUserMsgIndex !== undefined && secondLastMsgUserIndex !== undefined
+	clineMessages: Array<ClineStorageMessage | Anthropic.MessageParam>,
+	supportCache: boolean,
+): Array<Anthropic.MessageParam> {
+	// The latest message will be the new user message, one before will be the assistant message from a previous request,
+	// and the user message before that will be a previously cached user message. So we need to mark the latest user message
+	// as ephemeral to cache it for the next request, and mark the second to last user message as ephemeral to let the server
+	// know the last message to retrieve from the cache for the current request.
+	const userMsgIndices = clineMessages.reduce((acc, msg, index) => {
+		if (msg.role === "user") {
+			acc.push(index)
+		}
+		return acc
+	}, [] as number[])
+	// Set to -1 if there are no user messages so the indices are invalid
+	const indicesLength = userMsgIndices.length ?? -1
+	const lastUserMsgIndex = userMsgIndices[indicesLength - 1]
+	const secondLastMsgUserIndex = userMsgIndices[indicesLength - 2]
 
-		if (addCacheControl && (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 clineMessages.map((msg, index) => {
+		const anthropicMsg = convertClineStorageToAnthropicMessage(msg)
+
+		// Add cache control to the last two user messages
+		if (supportCache && (index === lastUserMsgIndex || index === secondLastMsgUserIndex)) {
+			return addCacheControl(anthropicMsg)
 		}
 
-		return {
-			...message,
-			content:
-				typeof message.content === "string"
-					? [
-							{
-								type: "text",
-								text: message.content,
-							},
-						]
-					: message.content,
-		}
+		return anthropicMsg
 	})
 }
 
+const isThinkingBlock = (
+	block: Anthropic.ContentBlockParam,
+): block is Anthropic.Messages.ThinkingBlockParam | Anthropic.Messages.RedactedThinkingBlockParam => {
+	return block.type === "thinking" || block.type === "redacted_thinking"
+}
+
 /**
- * Remove reasoning details from a single Anthropic message parameter
+ * Adds ephemeral cache control to the last content block of a message.
+ * Returns a new message object without mutating the original.
+ *
+ * @param message - The Anthropic message to add cache control to
+ * @returns A new message with cache control added to the last content block
  */
-function removeReasoningDetails(param: MessageParam): MessageParam {
-	if (Array.isArray(param.content)) {
+function addCacheControl(message: Anthropic.MessageParam): Anthropic.MessageParam {
+	// Convert string content to array format
+	if (typeof message.content === "string") {
 		return {
-			...param,
-			content: param.content.map((item) => {
-				if (item.type === "text") {
-					return {
-						...item,
-						reasoning_details: undefined,
-					}
-				}
-				return item
-			}),
+			...message,
+			content: [
+				{
+					type: "text",
+					text: message.content,
+					cache_control: { type: "ephemeral" },
+				} satisfies Anthropic.TextBlockParam,
+			],
 		}
 	}
-	return param
+
+	// Handle array content - add cache control to the last block
+	const content = [...message.content]
+	const lastIndex = content.length - 1
+
+	if (lastIndex >= 0) {
+		const lastBlock = content[lastIndex]
+
+		// Only add cache_control to block types that support it (not ThinkingBlockParam)
+		if (!isThinkingBlock(lastBlock)) {
+			content[lastIndex] = {
+				...lastBlock,
+				cache_control: { type: "ephemeral" },
+			} satisfies Anthropic.ContentBlockParam
+		}
+	}
+
+	return { ...message, content }
 }

src/core/api/transform/gemini-format.ts

@@ -1,7 +1,8 @@
 import { Anthropic } from "@anthropic-ai/sdk"
 import { Content, GenerateContentResponse, Part } from "@google/genai"
+import { ClineStorageMessage } from "@/shared/messages/content"
 
-export function convertAnthropicContentToGemini(content: string | Anthropic.ContentBlockParam[]): Part[] {
+export function convertAnthropicContentToGemini(content: string | ClineStorageMessage["content"]): Part[] {
 	if (typeof content === "string") {
 		return [{ text: content }]
 	}
@@ -9,7 +10,7 @@ export function convertAnthropicContentToGemini(content: string | Anthropic.Cont
 		.flatMap((block): Part | undefined => {
 			switch (block.type) {
 				case "text":
-					return { text: block.text }
+					return { text: block.text, thoughtSignature: block.signature }
 				case "image":
 					if (block.source.type !== "base64") {
 						throw new Error("Unsupported image source type")
@@ -26,6 +27,7 @@ export function convertAnthropicContentToGemini(content: string | Anthropic.Cont
 							name: block.name,
 							args: block.input as Record<string, unknown>,
 						},
+						thoughtSignature: block.signature,
 					}
 				case "tool_result":
 					return {

src/core/api/transform/ollama-format.ts

@@ -1,7 +1,13 @@
-import { Anthropic } from "@anthropic-ai/sdk"
 import { Message } from "ollama"
+import {
+	ClineAssistantToolUseBlock,
+	ClineImageContentBlock,
+	ClineStorageMessage,
+	ClineTextContentBlock,
+	ClineUserToolResultContentBlock,
+} from "@/shared/messages/content"
 
-export function convertToOllamaMessages(anthropicMessages: Anthropic.Messages.MessageParam[]): Message[] {
+export function convertToOllamaMessages(anthropicMessages: Omit<ClineStorageMessage, "modelInfo">[]): Message[] {
 	const ollamaMessages: Message[] = []
 
 	for (const anthropicMessage of anthropicMessages) {
@@ -13,8 +19,8 @@ export function convertToOllamaMessages(anthropicMessages: Anthropic.Messages.Me
 		} else {
 			if (anthropicMessage.role === "user") {
 				const { nonToolMessages, toolMessages } = anthropicMessage.content.reduce<{
-					nonToolMessages: (Anthropic.TextBlockParam | Anthropic.ImageBlockParam)[]
-					toolMessages: Anthropic.ToolResultBlockParam[]
+					nonToolMessages: (ClineTextContentBlock | ClineImageContentBlock)[]
+					toolMessages: ClineUserToolResultContentBlock[]
 				}>(
 					(acc, part) => {
 						if (part.type === "tool_result") {
@@ -70,8 +76,8 @@ export function convertToOllamaMessages(anthropicMessages: Anthropic.Messages.Me
 				}
 			} else if (anthropicMessage.role === "assistant") {
 				const { nonToolMessages, toolMessages } = anthropicMessage.content.reduce<{
-					nonToolMessages: (Anthropic.TextBlockParam | Anthropic.ImageBlockParam)[]
-					toolMessages: Anthropic.ToolUseBlockParam[]
+					nonToolMessages: (ClineTextContentBlock | ClineImageContentBlock)[]
+					toolMessages: ClineAssistantToolUseBlock[]
 				}>(
 					(acc, part) => {
 						if (part.type === "tool_use") {

src/core/api/transform/openai-format.ts

@@ -1,8 +1,26 @@
 import { Anthropic } from "@anthropic-ai/sdk"
 import OpenAI from "openai"
+import {
+	ClineAssistantRedactedThinkingBlock,
+	ClineAssistantThinkingBlock,
+	ClineAssistantToolUseBlock,
+	ClineImageContentBlock,
+	ClineStorageMessage,
+	ClineTextContentBlock,
+	ClineUserToolResultContentBlock,
+} from "@/shared/messages/content"
 
+/**
+ * Converts an array of ClineStorageMessage objects to OpenAI's Completions API format.
+ *
+ * Handles conversion of Cline-specific content types (tool uses, tool results, images, reasoning details)
+ * into OpenAI's expected message structure, including tool_calls and tool_call_id fields.
+ *
+ * @param anthropicMessages - Array of ClineStorageMessage objects to be converted
+ * @returns Array of OpenAI.Chat.ChatCompletionMessageParam objects
+ */
 export function convertToOpenAiMessages(
-	anthropicMessages: Anthropic.Messages.MessageParam[],
+	anthropicMessages: Omit<ClineStorageMessage, "modelInfo">[],
 ): OpenAI.Chat.ChatCompletionMessageParam[] {
 	const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = []
 
@@ -23,8 +41,8 @@ export function convertToOpenAiMessages(
          */
 			if (anthropicMessage.role === "user") {
 				const { nonToolMessages, toolMessages } = anthropicMessage.content.reduce<{
-					nonToolMessages: (Anthropic.TextBlockParam | Anthropic.ImageBlockParam)[]
-					toolMessages: Anthropic.ToolResultBlockParam[]
+					nonToolMessages: (ClineTextContentBlock | ClineImageContentBlock)[]
+					toolMessages: ClineUserToolResultContentBlock[]
 				}>(
 					(acc, part) => {
 						if (part.type === "tool_result") {
@@ -38,7 +56,7 @@ export function convertToOpenAiMessages(
 				)
 
 				// Process tool result messages FIRST since they must follow the tool use messages
-				const toolResultImages: Anthropic.Messages.ImageBlockParam[] = []
+				const toolResultImages: ClineImageContentBlock[] = []
 				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 OpenAI 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
@@ -102,8 +120,13 @@ export function convertToOpenAiMessages(
 				}
 			} else if (anthropicMessage.role === "assistant") {
 				const { nonToolMessages, toolMessages } = anthropicMessage.content.reduce<{
-					nonToolMessages: (Anthropic.TextBlockParam | Anthropic.ImageBlockParam)[]
-					toolMessages: Anthropic.ToolUseBlockParam[]
+					nonToolMessages: (
+						| ClineTextContentBlock
+						| ClineImageContentBlock
+						| ClineAssistantThinkingBlock
+						| ClineAssistantRedactedThinkingBlock
+					)[]
+					toolMessages: ClineAssistantToolUseBlock[]
 				}>(
 					(acc, part) => {
 						if (part.type === "tool_use") {
@@ -119,6 +142,7 @@ export function convertToOpenAiMessages(
 				// Process non-tool messages
 				let content: string | undefined
 				const reasoningDetails: any[] = []
+				const thinkingBlock = []
 				if (nonToolMessages.length > 0) {
 					nonToolMessages.forEach((part) => {
 						// @ts-ignore-next-line
@@ -134,27 +158,42 @@ export function convertToOpenAiMessages(
 							// @ts-ignore-next-line
 							// delete part.reasoning_details
 						}
+						if (part.type === "thinking" && part.thinking) {
+							// Reasoning details should have been moved to the text block
+							thinkingBlock.push(part)
+						}
 					})
 					content = nonToolMessages
 						.map((part) => {
-							if (part.type === "image") {
-								return "" // impossible as the assistant cannot send images
+							if (part.type === "text" && part.text) {
+								return part.text
 							}
-							return part.text
+							return ""
 						})
 						.join("\n")
 				}
 
 				// Process tool use messages
-				const tool_calls: OpenAI.Chat.ChatCompletionMessageToolCall[] = toolMessages.map((toolMessage) => ({
-					id: toolMessage.id,
-					type: "function",
-					function: {
-						name: toolMessage.name,
-						// json string
-						arguments: JSON.stringify(toolMessage.input),
-					},
-				}))
+				const tool_calls: OpenAI.Chat.ChatCompletionMessageToolCall[] = toolMessages.map((toolMessage) => {
+					const toolDetails = toolMessage.reasoning_details
+					if (toolDetails?.length) {
+						if (Array.isArray(toolDetails)) {
+							reasoningDetails.push(...toolDetails)
+						} else {
+							reasoningDetails.push(toolDetails)
+						}
+					}
+
+					return {
+						id: toolMessage.id,
+						type: "function",
+						function: {
+							name: toolMessage.name,
+							// json string
+							arguments: JSON.stringify(toolMessage.input),
+						},
+					}
+				})
 
 				// Set content to blank when tool_calls are present but content has no text, per OpenAI API spec
 				const hasToolCalls = tool_calls.length > 0
@@ -321,29 +360,30 @@ export function convertToAnthropicMessage(completion: OpenAI.Chat.Completions.Ch
 	}
 	try {
 		if (openAiMessage?.tool_calls?.length) {
-			anthropicMessage.content.push(
-				...openAiMessage.tool_calls
-					.map((toolCall): Anthropic.ToolUseBlock => {
-						const parsedName = toolCall.type === "function" && toolCall.function.name
-						let parsedInput = toolCall.function.arguments
+			const functionCalls = openAiMessage.tool_calls.filter((tc: any) => tc?.type === "function" && tc.function)
+			if (functionCalls.length > 0) {
+				anthropicMessage.content.push(
+					...functionCalls.map((toolCall: any): Anthropic.ToolUseBlock => {
+						let parsedInput = {}
 						try {
-							parsedInput = JSON.parse(toolCall.function.arguments || "{}")
+							parsedInput = JSON.parse(toolCall.function?.arguments || "{}")
 						} catch (error) {
 							console.error("Failed to parse tool arguments:", error)
 						}
 						return {
 							type: "tool_use",
 							id: toolCall.id,
-							name: parsedName || UNIQUE_ERROR_TOOL_NAME,
+							name: toolCall.function?.name || UNIQUE_ERROR_TOOL_NAME,
 							input: parsedInput,
 						}
-					})
-					// Filter out any tool uses with the UNIQUE_ERROR_TOOL_NAME, which indicates a parsing error
-					.filter((toolUse) => toolUse.name !== UNIQUE_ERROR_TOOL_NAME),
-			)
+					}),
+				)
+			}
+
+			return anthropicMessage
 		}
 	} catch (error) {
-		console.error("Failed to process tool calls:", error)
+		console.error("Error converting OpenAI message to Anthropic format:", error)
 	}
 
 	return anthropicMessage

src/core/api/transform/openai-response-format.ts

@@ -0,0 +1,217 @@
+import { ResponseInput, ResponseInputMessageContentList, ResponseReasoningItem } from "openai/resources/responses/responses"
+import { ClineStorageMessage } from "@/shared/messages/content"
+
+/**
+ * Converts an array of ClineStorageMessage objects (extension of Anthropic format) to a ResponseInput array to use with OpenAI's Responses API.
+ *
+ * ## Key Differences from Chat Completions API
+ *
+ * The Responses API has stricter requirements than the Chat Completions API:
+ *
+ * ### Chat Completions API:
+ * - Messages are simple role/content pairs
+ * - System prompts are separate messages with role="system"
+ * - No explicit reasoning item structure
+ * - More forgiving about message ordering
+ *
+ * ### Responses API:
+ * - Uses an "input" array of heterogeneous items (messages, reasoning, function_calls, etc.)
+ * - System prompts go in an "instructions" field, not as messages
+ * - Reasoning items MUST be immediately followed by a message or function_call
+ * - Strict ordering requirements match training data distribution
+ *
+ * ## The Reasoning Item Constraint
+ *
+ * **THE CRITICAL ERROR:** "Item 'rs_...' of type 'reasoning' was provided without its required following item"
+ *
+ * This error occurs when reasoning items are orphaned or separated from their corresponding output.
+ *
+ * ### What Causes This Error:
+ * ```
+ * ❌ WRONG - Reasoning orphaned between turns:
+ * [
+ *   { role: "user", content: [...] },
+ *   { type: "reasoning", id: "rs_abc", summary: [...] },  // ← ORPHANED!
+ *   { type: "message", role: "assistant", content: [...] },
+ *   { role: "user", content: [...] }
+ * ]
+ * ```
+ *
+ * ### The Fix - Keep Complete Assistant Turns Together:
+ * ```
+ * ✅ CORRECT - Reasoning paired with its message:
+ * [
+ *   { role: "user", content: [...] },
+ *   { type: "reasoning", id: "rs_abc", summary: [...] },
+ *   { type: "message", role: "assistant", content: [...] },  // ← Immediately follows reasoning
+ *   { role: "user", content: [...] }
+ * ]
+ * ```
+ *
+ * **Per OpenAI Engineering Guidance:**
+ * - ❌ WRONG: `content += filter(lambda x: x.type == "reasoning", resp.output)`
+ * - ✅ CORRECT: `content += resp.output`
+ *
+ * Never extract only reasoning items - always include the complete output sequence
+ * (reasoning + message/function_call) as provided by the API.
+ *
+ * ## Implementation Strategy
+ *
+ * 1. **Separate processing for assistant vs user messages** - Assistant turns need special
+ *    handling to maintain reasoning-message pairing
+ * 2. **Collect all assistant items together** - Gather reasoning, messages, and function_calls
+ *    for the entire assistant turn before validating
+ * 3. **Validate pairing within each turn** - Ensure each reasoning item is immediately followed
+ *    by a message or function_call, inserting placeholders if needed
+ * 4. **Flush complete turns atomically** - Add all items from an assistant turn together to
+ *    maintain proper sequencing
+ *
+ * @link https://community.openai.com/t/openai-api-error-function-call-was-provided-without-its-required-reasoning-item-the-real-issue/1355347
+ *
+ * @param messages - Array of ClineStorageMessage objects to be converted
+ * @returns ResponseInput array containing the transformed messages with proper reasoning pairing
+ */
+export function convertToOpenAIResponsesInput(messages: ClineStorageMessage[]): ResponseInput {
+	const allItems: any[] = []
+	const toolUseIdToCallId = new Map<string, string>()
+
+	for (const m of messages) {
+		if (typeof m.content === "string") {
+			allItems.push({ role: m.role, content: [{ type: "input_text", text: m.content }] })
+			continue
+		}
+
+		if (m.role === "assistant") {
+			// For assistant messages, we must ensure reasoning items are IMMEDIATELY followed
+			// by their corresponding message or function_call. Process the entire assistant
+			// turn and ensure proper pairing.
+			const assistantItems: any[] = []
+
+			for (const part of m.content) {
+				switch (part.type) {
+					case "thinking":
+						// Include reasoning item if it has a call_id, even if thinking is empty
+						// This is required because the API expects reasoning items to be paired with
+						// their corresponding function_calls, and will error if a function_call
+						// references a reasoning item that wasn't sent
+						if (part.call_id && part.call_id.length > 0) {
+							assistantItems.push({
+								id: part.call_id,
+								type: "reasoning",
+								summary: part.thinking
+									? [
+											{
+												type: "summary_text",
+												text: part.thinking,
+											},
+										]
+									: [],
+							} as ResponseReasoningItem)
+						}
+						break
+					case "redacted_thinking":
+						// Include reasoning item with encrypted content if it has a call_id
+						// Even if data is missing, we need to maintain the reasoning-function_call pairing
+						if (part.call_id && part.call_id.length > 0) {
+							const reasoningItem: any = {
+								id: part.call_id,
+								type: "reasoning",
+								summary: [],
+							}
+							// Only include encrypted_content if data exists
+							if (part.data) {
+								reasoningItem.encrypted_content = part.data
+							}
+							assistantItems.push(reasoningItem as ResponseReasoningItem)
+						}
+						break
+					case "text":
+						assistantItems.push({
+							type: "message",
+							role: "assistant",
+							content: [{ type: "output_text", text: part.text }],
+						})
+						break
+					case "image":
+						assistantItems.push({
+							type: "message",
+							role: "assistant",
+							content: [{ type: "output_text", text: `[image:${part.source.media_type}]` }],
+						})
+						break
+					case "tool_use": {
+						const call_id = part.call_id || part.id
+						if (part.call_id) {
+							toolUseIdToCallId.set(part.id, part.call_id)
+						}
+						assistantItems.push({
+							type: "function_call",
+							call_id,
+							id: part.id,
+							name: part.name,
+							arguments: JSON.stringify(part.input ?? {}),
+						})
+						break
+					}
+				}
+			}
+
+			// Ensure every reasoning item is followed by a message or function_call
+			for (let i = 0; i < assistantItems.length; i++) {
+				const item = assistantItems[i]
+				if (item.type === "reasoning") {
+					const nextItem = assistantItems[i + 1]
+					if (!nextItem || (nextItem.type !== "message" && nextItem.type !== "function_call")) {
+						// Insert a placeholder message immediately after this reasoning item
+						assistantItems.splice(i + 1, 0, {
+							type: "message",
+							role: "assistant",
+							content: [{ type: "output_text", text: "" }],
+						})
+					}
+				}
+			}
+
+			allItems.push(...assistantItems)
+		} else {
+			// User messages - collect all content
+			const messageContent: ResponseInputMessageContentList = []
+
+			for (const part of m.content) {
+				switch (part.type) {
+					case "text":
+						messageContent.push({ type: "input_text", text: part.text })
+						break
+					case "image":
+						messageContent.push({
+							type: "input_image",
+							detail: "auto",
+							image_url: `data:${part.source.media_type};base64,${part.source.data}`,
+						})
+						break
+					case "tool_result": {
+						// Flush any pending message content before adding tool result
+						if (messageContent.length > 0) {
+							allItems.push({ role: m.role, content: [...messageContent] })
+							messageContent.length = 0
+						}
+						const call_id = part.call_id || toolUseIdToCallId.get(part.tool_use_id) || part.tool_use_id
+						allItems.push({
+							type: "function_call_output",
+							call_id,
+							output: typeof part.content === "string" ? part.content : JSON.stringify(part.content),
+						})
+						break
+					}
+				}
+			}
+
+			// Flush any remaining user message content
+			if (messageContent.length > 0) {
+				allItems.push({ role: m.role, content: [...messageContent] })
+			}
+		}
+	}
+
+	return allItems
+}