How GetRouter Works

GetRouter provides a single account, API key system, and public API surface for multiple AI model capabilities.

Your application or AI tool
            ↓
     GetRouter API
            ↓
   Selected model service
            ↓
        Response

Request flow

Create an API key. The key identifies and authorizes your requests.
Choose a model ID. Model IDs are listed on the live Models page.
Use the matching endpoint. For example, use /v1/chat/completions for Chat Completions or /v1/responses for the Responses API.
Send the request. Include your key as a Bearer token and pass the model ID in the request.
Read the response. Synchronous endpoints return a response directly; streaming endpoints send incremental events; video endpoints can return a task that you poll.
Review usage. Usage and charges are recorded in your GetRouter account according to the selected model and request.

A valid model ID does not mean the model supports every endpoint. Match the model to the protocol and capability required by your application:

Use case	Common endpoint
OpenAI-compatible chat	`POST /v1/chat/completions`
Responses-based agents and coding tools	`POST /v1/responses`
Claude-compatible clients	`POST /v1/messages`
Gemini-compatible clients	`POST /v1beta/models/{model}:generateContent`
Text embeddings	`POST /v1/embeddings`
Image generation	`POST /v1/images/generations`
Speech generation	`POST /v1/audio/speech`
Video generation	`POST /v1/videos`

Tip

If an integration returns model not found or an endpoint compatibility error, verify both the exact model ID and the protocol used by the tool.