diff --git a/docs/docs.json b/docs/docs.json index 665874f8..2bead952 100644 --- a/docs/docs.json +++ b/docs/docs.json @@ -41,6 +41,7 @@ "llm-providers/bedrock", "llm-providers/azure", "llm-providers/novita", + "llm-providers/anonrouter", "llm-providers/local" ] }, diff --git a/docs/llm-providers/anonrouter.mdx b/docs/llm-providers/anonrouter.mdx new file mode 100644 index 00000000..62dfab85 --- /dev/null +++ b/docs/llm-providers/anonrouter.mdx @@ -0,0 +1,39 @@ +--- +title: "AnonRouter" +description: "Configure Strix with models via AnonRouter" +--- + +[AnonRouter](https://anonrouter.ai) provides an OpenAI-compatible API for models from multiple providers. + +## Setup + +Create an AnonRouter API key, then configure Strix: + +```bash +export STRIX_LLM="openai/" +export LLM_API_KEY="your-anonrouter-api-key" +export LLM_API_BASE="https://api.anonrouter.ai/v1" +``` + +The first `openai/` segment tells Strix to use its OpenAI-compatible client. The remaining value is the AnonRouter model ID, passed through unchanged — keep its provider prefix and capitalization exactly as AnonRouter lists it. + +Setting `LLM_API_BASE` also makes Strix use the chat completions API rather than the responses API, which is the path AnonRouter serves. + +## Available Models + +Use any model returned by the AnonRouter models endpoint, which requires your API key: + +```bash +curl https://api.anonrouter.ai/v1/models \ + -H "Authorization: Bearer $LLM_API_KEY" +``` + +Prefix the model ID with `openai/` when setting `STRIX_LLM`, keeping the rest of the ID byte for byte. A listed model ID of the form `/` becomes `openai//` in Strix; a listed ID with no slash becomes `openai/`. + +Strix drives scans with streamed tool calls, so pick a model that supports function calling. + +## Get API Key + +1. Sign up at [anonrouter.ai](https://anonrouter.ai) +2. Create an API key +3. Set the key as `LLM_API_KEY`