Introduction
GetRouter gives you one API key and a unified set of endpoints for accessing multiple AI models. You can use it from your own application, an OpenAI-compatible SDK, or a supported coding and chat tool.
GetRouter is designed for two common workflows:
- Build with the API — call text, embedding, image, audio, video, rerank, Claude-compatible, and Gemini-compatible endpoints from your code.
- Connect an AI tool — enter a GetRouter API key, Base URL, and model ID in tools such as Codex, OpenClaw, CC Switch, and OpenAI-compatible clients.
What you need
Every integration starts with three values:
Requests use Bearer authentication:
Keep your API key on the server or in the tool's protected credential store. Do not publish it in source code, screenshots, browser-side JavaScript, or support messages.
Supported API styles
GetRouter exposes several API styles so you can use the protocol expected by your application or tool:
- OpenAI-compatible Chat Completions, Responses, Embeddings, Images, Audio, and Models endpoints
- Anthropic Claude-compatible Messages endpoint
- Gemini-compatible Models and
generateContentendpoints - Asynchronous video generation endpoints
- Rerank endpoint
Protocol support and request fields can vary by model. Choose a model that supports the endpoint and capability you intend to use.
Choose your path
Build with the API
Start with API Quick Start, then continue to API Guides or the API Reference.
Connect a tool
Start with Tool Quick Start, then follow the guide for your tool under Integrations.
Live model and pricing information
Model availability and pricing can change. Use these live pages instead of relying on a copied model list:
