ModRouter is an AI gateway for routing inference requests across configured model providers. It gives applications one place to call chat, responses, embeddings, images, audio, files, batches, and realtime APIs.

​
What You Can Do

Use AI inference APIs

Call OpenAI-compatible inference endpoints and provider-compatible SDK routes.

List available models

Discover the models exposed by your ModRouter instance.

Use MCP tools

Connect external tools through MCP and expose them to compatible clients.

Use existing SDKs

Keep OpenAI, Anthropic, Bedrock, GenAI, LiteLLM, LangChain, or PydanticAI clients.

​
API Shapes

ModRouter exposes:
  • OpenAI-compatible /v1/* inference endpoints.
  • Provider SDK compatibility routes such as /openai, /anthropic, /bedrock, and /genai.
  • MCP endpoints for tool access.
  • A model listing endpoint for discovering available models.