Skip to content
Router One

CONNECT YOUR TOOLS

Your tools. One API key.

Find your tool, follow its setup guide, and verify your first model call. Router One handles the model request; your application stays in charge of its agents and workflows.

Find your integration

Choose what you use, then check its provider settings and model requirements.

Check API addresses

38 of 38 integration guides

Coding & agent tools

16

SDKs & frameworks

12

Chat, RAG & translation

06

Workflows & API tools

04

Before your first request

  1. Match the URL to your client

    A base URL and a full request endpoint are different. Check the client's field and selected protocol before copying an address.

    Check connection settings
  2. Use an exact model ID

    Copy an ID from the model catalog. Match its supported endpoints and capabilities to the features your client needs.

    Browse models and capabilities
  3. Verify the finished request

    Start with plain text. Check the result, request ID, and settlement in Logs before enabling streaming or tools.

    Read the request log guide

Which URL should I enter?

OpenAI-compatible SDKs and Codex use the /v1 base URL; Claude Code uses the host root; Immersive Translate needs the full /v1/chat/completions URL. Check the configured protocol and model support.

Compare client URL settings
  • SDK Chat Completions setupbase_url / baseURL / apiBase
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • Codex CLIbase_url; wire_api = "responses"
    Value to enter
    https://api.router.one/v1
    API used
    ResponsesPOST /v1/responses
  • Claude CodeANTHROPIC_BASE_URL
    Value to enter
    https://api.router.one
    API used
    Anthropic MessagesPOST /v1/messages
  • Immersive TranslateCustom endpoint
    Value to enter
    https://api.router.one/v1/chat/completions
    API used
    Chat CompletionsPOST /v1/chat/completions
  • GitHub Copilot (VS Code)chatLanguageModels.json → models[].url; apiType: "chat-completions"
    Value to enter
    https://api.router.one/v1/chat/completions
    API used
    Chat CompletionsPOST /v1/chat/completions
  • OpenHandsBase URL (Advanced) / LLM_BASE_URL (GPT-5-family IDs use /v1/responses; see the guide)
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • NextChatBASE_URL / OpenAI Endpoint
    Value to enter
    https://api.router.one
    API used
    Chat CompletionsPOST /v1/chat/completions
  • CrewAIbase_url; custom_openai=True (or OPENAI_API_BASE)
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • Spring AIspring.ai.openai.base-url (2.0.x; 1.x uses the host root)
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • OpenAI Agents SDKAsyncOpenAI(base_url=…) + OpenAIChatCompletionsModel (string model names default to /v1/responses)
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • Claude Agent SDKClaudeAgentOptions(env={ANTHROPIC_BASE_URL, ANTHROPIC_AUTH_TOKEN})
    Value to enter
    https://api.router.one
    API used
    Anthropic MessagesPOST /v1/messages
  • HaystackOpenAIChatGenerator(api_base_url=…)
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • AnythingLLMSettings → LLM → Base URL / GENERIC_OPEN_AI_BASE_PATH
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • FastGPTAI Proxy channel → Base url (Protocol Type: OpenAI)
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • XcodeSettings → Intelligence → Add a Chat Provider → Internet Hosted → URL (Xcode appends /v1/models and /v1/chat/completions)
    Value to enter
    https://api.router.one
    API used
    Chat CompletionsPOST /v1/chat/completions
  • Qwen CodeOPENAI_BASE_URL (.qwen/.env) / modelProviders.openai[].baseUrl
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • Google ADKLiteLlm(model="openai/<id>", api_base=…, api_key=…)
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • smolagentsOpenAIModel(model_id=…, api_base=…, api_key=…)
    Value to enter
    https://api.router.one/v1
    API used
    Chat CompletionsPOST /v1/chat/completions
  • GooseOPENAI_HOST (OPENAI_BASE_PATH keeps its default v1/chat/completions; gpt-5, gpt-6 and o-series names go to /v1/responses, see the guide)
    Value to enter
    https://api.router.one
    API used
    Chat CompletionsPOST /v1/chat/completions

These are the configurations documented in the linked guides. A client can support other providers or APIs; verify the actual request path before changing the URL.

Integration questions, answered

Which Router One base URL should I use?

For the OpenAI-compatible SDK configuration in these guides, use https://api.router.one/v1. Claude Code and Claude Agent SDK use the host root, https://api.router.one. Clients that ask for a full Chat Completions endpoint need https://api.router.one/v1/chat/completions. Some clients append /v1 themselves; follow the guide for the actual field rather than adding it twice.

Does a listed integration support every model and feature?

A guide documents a particular provider configuration. The selected API, model, and client must all support the feature you need. Check the exact model ID, endpoints, context limit, and tool or image support on the model's detail page. A working text request does not verify streaming, tools, or image input.

Can Router One run my agent or index my knowledge base?

Your client or framework runs the agent loop, executes tools, and manages documents. Router One handles each supported model call. Configure embeddings, reranking, and vector storage separately when your knowledge base needs them; the chat connection alone does not provide those services.

How do I check that my integration worked?

Send a simple request and check its final result. In Dashboard → Logs, match the model, time, and request ID, then inspect tokens, latency, status, and settlement. HTTP 200 alone does not prove a stream completed, and pending settlement does not mean a request was free. The customer log view does not display provider attempts or the internal fallback chain.

Your tool isn't listed?

Check whether it accepts a custom provider URL and API key, then match its request protocol to a supported model. The API reference describes the available endpoints.

Read as Markdown