diff --git a/packages/kilo-docs/lib/nav/ai-providers.ts b/packages/kilo-docs/lib/nav/ai-providers.ts index bb9f109ae7..92ad10f650 100644 --- a/packages/kilo-docs/lib/nav/ai-providers.ts +++ b/packages/kilo-docs/lib/nav/ai-providers.ts @@ -42,6 +42,8 @@ export const AiProvidersNav: NavSection[] = [ links: [ { href: "/ai-providers/vertex", children: "Google Vertex AI" }, { href: "/ai-providers/bedrock", children: "AWS Bedrock" }, + { href: "/ai-providers/alibaba", children: "Alibaba Cloud" }, + { href: "/ai-providers/cloudflare", children: "Cloudflare" }, { href: "/ai-providers/groq", children: "Groq" }, { href: "/ai-providers/cerebras", children: "Cerebras" }, { href: "/ai-providers/fireworks", children: "Fireworks AI" }, @@ -68,6 +70,7 @@ export const AiProvidersNav: NavSection[] = [ { href: "/ai-providers/moonshot", children: "Moonshot" }, { href: "/ai-providers/ovhcloud", children: "OVHcloud" }, { href: "/ai-providers/sap-ai-core", children: "SAP AI Core" }, + { href: "/ai-providers/venice", children: "Venice AI" }, ], }, { diff --git a/packages/kilo-docs/pages/ai-providers/alibaba.md b/packages/kilo-docs/pages/ai-providers/alibaba.md new file mode 100644 index 0000000000..63ac703840 --- /dev/null +++ b/packages/kilo-docs/pages/ai-providers/alibaba.md @@ -0,0 +1,79 @@ +--- +sidebar_label: Alibaba Cloud +--- + +# Using Alibaba Cloud With Kilo Code + +Kilo Code supports Alibaba Cloud Model Studio (DashScope) through the native Alibaba AI SDK provider. Use it to run Qwen and other DashScope-hosted models directly from Kilo. + +**Website:** [https://www.alibabacloud.com/product/modelstudio](https://www.alibabacloud.com/product/modelstudio) + +## Getting an API Key + +1. Sign in to Alibaba Cloud Model Studio or DashScope. +2. Open the API key section for your workspace. +3. Create a key and copy it immediately. +4. Store the key securely; do not commit it to your repository. + +## Configuration in Kilo Code + +{% tabs %} +{% tab label="VSCode (Legacy)" %} + +Use the **OpenAI Compatible** provider if the legacy provider list does not include Alibaba Cloud. Set the base URL and API key from your DashScope account, then choose or enter a supported model ID. + +{% /tab %} +{% tab label="VSCode" %} + +Open **Settings** (gear icon) and go to the **Providers** tab to add Alibaba Cloud. Enter your DashScope API key and choose a supported model from the model picker. + +The extension stores this in your `kilo.json` config file. You can also edit the config file directly — see the **CLI** tab for the file format. + +{% /tab %} +{% tab label="CLI" %} + +Set the DashScope API key as an environment variable or configure it in your `kilo.json` config file: + +**Environment variable:** + +```bash +export DASHSCOPE_API_KEY="your-api-key" +``` + +**Config file** (`~/.config/kilo/kilo.json` or `./kilo.json`): + +```jsonc +{ + "provider": { + "alibaba": { + "env": ["DASHSCOPE_API_KEY"] + } + } +} +``` + +Then set your default model: + +```jsonc +{ + "model": "alibaba/qwen-plus" +} +``` + +Some regions and model catalogs may expose Alibaba models under `alibaba-cn` instead. In that case, use the same API key and set the model as `alibaba-cn/`. + +{% /tab %} +{% /tabs %} + +## Tips and Notes + +- **Native provider:** Kilo uses the native Alibaba AI SDK provider, not a generic compatibility shim, when you select the built-in Alibaba provider. +- **Cache controls:** Kilo forwards provider cache controls for compatible Alibaba/DashScope models, which can reduce repeated context cost where the model supports caching. +- **Regional catalogs:** Model availability varies by account, region, and gateway. Check the DashScope console for the exact model IDs available to your account. +- **Pricing and limits:** Refer to Alibaba Cloud Model Studio pricing and rate-limit documentation for current usage details. + +## Troubleshooting + +- **Authentication errors:** Verify `DASHSCOPE_API_KEY` is set in the same shell that launches Kilo, or stored in your Kilo provider config. +- **Model not found:** Confirm the model ID is available in your DashScope account and try the model picker to see Kilo's current catalog. +- **Region mismatch:** If your account uses the China-region catalog, try the `alibaba-cn/` provider prefix. diff --git a/packages/kilo-docs/pages/ai-providers/cloudflare.md b/packages/kilo-docs/pages/ai-providers/cloudflare.md new file mode 100644 index 0000000000..c7bb94a236 --- /dev/null +++ b/packages/kilo-docs/pages/ai-providers/cloudflare.md @@ -0,0 +1,128 @@ +--- +sidebar_label: Cloudflare +--- + +# Using Cloudflare With Kilo Code + +Kilo Code supports both Cloudflare Workers AI and Cloudflare AI Gateway. Workers AI runs Cloudflare-hosted models directly, while AI Gateway routes requests through your Cloudflare gateway to one or more upstream providers. + +**Website:** [https://developers.cloudflare.com/ai/](https://developers.cloudflare.com/ai/) + +## Provider Options + +| Provider ID | Use it for | Required values | +|---|---|---| +| `cloudflare-workers-ai` | Cloudflare-hosted Workers AI models | Account ID and API key | +| `cloudflare-ai-gateway` | Routing through Cloudflare AI Gateway | Account ID, Gateway ID, and Gateway API token | + +You can enter these values interactively with `/connect` in the TUI or `kilo auth` from the CLI, or provide them through environment variables. + +## Cloudflare Workers AI + +### Get Credentials + +1. Open the Cloudflare dashboard and select your account. +2. Copy the **Account ID** from the dashboard. +3. Create an API token that can invoke Workers AI. + +### Configure Kilo + +{% tabs %} +{% tab label="VSCode" %} + +Open **Settings** (gear icon) and go to the **Providers** tab to add Cloudflare Workers AI. If your account ID is not already available from the environment, Kilo prompts for it while connecting the provider. + +{% /tab %} +{% tab label="CLI" %} + +Use `/connect` in the TUI and choose Cloudflare Workers AI, or run: + +```bash +kilo auth cloudflare-workers-ai +``` + +Alternatively, set environment variables before launching Kilo: + +```bash +export CLOUDFLARE_ACCOUNT_ID="your-account-id" +export CLOUDFLARE_API_KEY="your-workers-ai-api-key" +``` + +Then choose a Workers AI model from the model picker, or set a default model such as: + +```jsonc +{ + "model": "cloudflare-workers-ai/@cf/meta/llama-3.1-8b-instruct" +} +``` + +{% /tab %} +{% /tabs %} + +## Cloudflare AI Gateway + +### Get Credentials + +1. Create or open a Cloudflare AI Gateway. +2. Copy your **Account ID** and **Gateway ID**. +3. Create a Gateway API token. + +### Configure Kilo + +{% tabs %} +{% tab label="VSCode" %} + +Open **Settings** (gear icon) and go to the **Providers** tab to add Cloudflare AI Gateway. Kilo prompts for the account ID and gateway ID when they are not already set in the environment. + +{% /tab %} +{% tab label="CLI" %} + +Use `/connect` in the TUI and choose Cloudflare AI Gateway, or run: + +```bash +kilo auth cloudflare-ai-gateway +``` + +Alternatively, set environment variables before launching Kilo: + +```bash +export CLOUDFLARE_ACCOUNT_ID="your-account-id" +export CLOUDFLARE_GATEWAY_ID="your-gateway-id" +export CLOUDFLARE_API_TOKEN="your-gateway-api-token" +``` + +`CF_AIG_TOKEN` is also accepted as an alternative to `CLOUDFLARE_API_TOKEN`. + +Then choose a gateway-backed model from the model picker, or set a default model such as: + +```jsonc +{ + "model": "cloudflare-ai-gateway/openai/gpt-5.1" +} +``` + +{% /tab %} +{% /tabs %} + +## Advanced Configuration + +If your organization provides a fully configured gateway URL, set `provider..options.baseURL`. When `baseURL` is configured, Kilo skips the built-in Account ID and Gateway ID checks because the URL already identifies the target service. + +```jsonc +{ + "provider": { + "cloudflare-ai-gateway": { + "options": { + "baseURL": "https://gateway.example.com/v1" + } + } + } +} +``` + +## Troubleshooting + +- **`CLOUDFLARE_ACCOUNT_ID is missing`** — set `CLOUDFLARE_ACCOUNT_ID` or reconnect the provider with `/connect` so Kilo can store the account ID in auth metadata. +- **`CLOUDFLARE_ACCOUNT_ID and CLOUDFLARE_GATEWAY_ID missing`** — set both values or run `kilo auth cloudflare-ai-gateway`. +- **`CLOUDFLARE_API_TOKEN (or CF_AIG_TOKEN) is required`** — set one of those variables or reconnect Cloudflare AI Gateway with `kilo auth cloudflare-ai-gateway`. +- **Model errors through AI Gateway** — confirm the gateway route supports the selected upstream provider/model and that your Gateway API token has access. diff --git a/packages/kilo-docs/pages/ai-providers/index.md b/packages/kilo-docs/pages/ai-providers/index.md index bc9822b1db..a7c8c05fee 100644 --- a/packages/kilo-docs/pages/ai-providers/index.md +++ b/packages/kilo-docs/pages/ai-providers/index.md @@ -22,6 +22,10 @@ Major AI companies offering powerful models via API: - **[Anthropic](/docs/ai-providers/anthropic)** - Claude models (Claude 4, Claude 3.5 Sonnet, etc.) - **[OpenAI](/docs/ai-providers/openai)** - GPT-4, GPT-4o, o1, and more - **[Google Gemini](/docs/ai-providers/gemini)** - Gemini Pro, Gemini Ultra +- **[Google Vertex AI](/docs/ai-providers/vertex)** - Google Cloud-hosted Gemini and partner models +- **[AWS Bedrock](/docs/ai-providers/bedrock)** - AWS-hosted foundation models +- **[Alibaba Cloud](/docs/ai-providers/alibaba)** - DashScope and Qwen models through Model Studio +- **[Cloudflare](/docs/ai-providers/cloudflare)** - Workers AI and Cloudflare AI Gateway - **[DeepSeek](/docs/ai-providers/deepseek)** - DeepSeek V3., R1 - **[Mistral](/docs/ai-providers/mistral)** - Mistral Large, Codestral @@ -40,6 +44,7 @@ Route requests through unified APIs with additional features: - **[OpenRouter](/docs/ai-providers/openrouter)** - Access multiple providers through one API - **[Glama](/docs/ai-providers/glama)** - Enterprise AI gateway - **[Requesty](/docs/ai-providers/requesty)** - Smart routing and fallbacks +- **[Cloudflare AI Gateway](/docs/ai-providers/cloudflare)** - Route providers through your Cloudflare account ## Choosing a Provider diff --git a/packages/kilo-docs/pages/ai-providers/moonshot.md b/packages/kilo-docs/pages/ai-providers/moonshot.md index 7b92cf1028..463dc1bb59 100644 --- a/packages/kilo-docs/pages/ai-providers/moonshot.md +++ b/packages/kilo-docs/pages/ai-providers/moonshot.md @@ -70,5 +70,6 @@ Then set your default model: - **Ultra-Long Context:** Kimi models excel at handling large codebases and complex projects with their extended context windows. - **Reasoning Capabilities:** The K2-Thinking variant provides enhanced problem-solving through extended reasoning chains. +- **Kimi-specific prompting:** Kilo automatically uses a Kimi-tuned system prompt for model IDs containing `kimi`; no extra configuration is required. - **Language Support:** Kimi models have strong support for both English and Chinese languages. - **Pricing:** Refer to the Moonshot.ai platform for current pricing information on different models. diff --git a/packages/kilo-docs/pages/ai-providers/venice.md b/packages/kilo-docs/pages/ai-providers/venice.md new file mode 100644 index 0000000000..2c45bb6032 --- /dev/null +++ b/packages/kilo-docs/pages/ai-providers/venice.md @@ -0,0 +1,76 @@ +--- +sidebar_label: Venice AI +--- + +# Using Venice AI With Kilo Code + +Kilo Code supports Venice AI through the native Venice provider. Venice offers privacy-focused access to open and reasoning-capable models through its API. + +**Website:** [https://venice.ai/](https://venice.ai/) + +## Getting an API Key + +1. Sign in to Venice AI. +2. Open your API settings. +3. Create a key for Kilo Code. +4. Copy the key immediately and store it securely. + +## Configuration in Kilo Code + +{% tabs %} +{% tab label="VSCode (Legacy)" %} + +Use the **OpenAI Compatible** provider if the legacy provider list does not include Venice AI. Enter your Venice API base URL, API key, and model ID from the Venice dashboard. + +{% /tab %} +{% tab label="VSCode" %} + +Open **Settings** (gear icon) and go to the **Providers** tab to add Venice AI and enter your API key. + +The extension stores this in your `kilo.json` config file. You can also edit the config file directly — see the **CLI** tab for the file format. + +{% /tab %} +{% tab label="CLI" %} + +Set the API key as an environment variable or configure it in your `kilo.json` config file: + +**Environment variable:** + +```bash +export VENICE_API_KEY="your-api-key" +``` + +**Config file** (`~/.config/kilo/kilo.json` or `./kilo.json`): + +```jsonc +{ + "provider": { + "venice": { + "env": ["VENICE_API_KEY"] + } + } +} +``` + +Then select a Venice model from the model picker, or set a default model after confirming the model ID in your account: + +```jsonc +{ + "model": "venice/" +} +``` + +{% /tab %} +{% /tabs %} + +## Tips and Notes + +- **Native provider:** Kilo uses the Venice AI SDK provider for built-in Venice models. +- **Reasoning models:** Some Venice models expose reasoning controls. Kilo maps supported reasoning effort and disable-thinking options when the selected model supports them. +- **Model availability:** Venice model IDs and access can vary by plan. Use the model picker or Venice dashboard for the current list. +- **Pricing:** Refer to Venice AI pricing and account limits for current token costs and rate limits. + +## Troubleshooting + +- **Invalid API key:** Verify `VENICE_API_KEY` is set in the same environment that launches Kilo, or reconnect the provider in Settings. +- **Model not available:** Confirm your Venice account has access to the selected model and try selecting it from the model picker. diff --git a/packages/kilo-docs/pages/ai-providers/vertex.md b/packages/kilo-docs/pages/ai-providers/vertex.md index abbb56da3b..26fef6fbcf 100644 --- a/packages/kilo-docs/pages/ai-providers/vertex.md +++ b/packages/kilo-docs/pages/ai-providers/vertex.md @@ -81,4 +81,5 @@ Then set your default model: ## Tips and Notes - **Permissions:** Ensure your Google Cloud account has the necessary permissions to access Vertex AI and the specific models you want to use. +- **Prompt caching:** Claude models served through Vertex AI support Kilo prompt caching. Kilo applies Anthropic cache controls and tracks cache write/read tokens when Vertex reports them. Native Vertex Gemini models use Google's implicit server-side caching; no extra Kilo configuration is required, and Gemini may not report cache write tokens. - **Pricing:** Refer to the [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) page for details. diff --git a/packages/kilo-docs/pages/customize/context/large-projects.md b/packages/kilo-docs/pages/customize/context/large-projects.md index 101fb1621b..cfa129e40b 100644 --- a/packages/kilo-docs/pages/customize/context/large-projects.md +++ b/packages/kilo-docs/pages/customize/context/large-projects.md @@ -31,7 +31,7 @@ The context window includes: 5. **Prioritize Recent History:** Kilo Code automatically truncates older messages in the conversation history to stay within the context window. Be mindful of this, and re-include important context if needed. -6. **Use Prompt Caching (if available):** Some API providers like Anthropic, OpenAI, OpenRouter and Requesty support "prompt caching". This caches your prompts for use in future tasks and helps reduce the cost and latency of requests. +6. **Use Prompt Caching (if available):** Some API providers like Anthropic, OpenAI, OpenRouter, Requesty, and Vertex AI Claude support "prompt caching". This caches your prompts for use in future tasks and helps reduce the cost and latency of requests. ## Example: Refactoring a Large File