1
Your App Sends a Request
Standard OpenAI format request with your chosen model ID, sent to the OpenRouter endpoint.
2
OpenRouter Routes to the Best Provider
The platform evaluates available providers for that model and routes to the fastest, cheapest, or most reliable one based on your preferences.
3
Provider Processes the Request
The underlying provider (OpenAI, Anthropic, Google, etc.) generates the response using the model you selected.
4
Response Returned in Standard Format
The response comes back in OpenAI compatible format, regardless of which provider actually processed it. Your app handles it the same way every time.
0
Models Available
0
API Compatibility
0
Switch Models
Migration tip for existing OpenAI users
If you already have code using the OpenAI SDK, migration takes about 30 seconds. Change the
base_url to https://openrouter.ai/api/v1, swap in your OpenRouter API key, and optionally change the model name. All existing parameters (temperature, max_tokens, streaming, tools) work the same way.