← AI solutions

AI services

OpenRouter

OpenRouter is a unified interface for calling AI models from many companies through a single API.

Overview

OpenRouter is a service for calling AI models from many providers through a single API endpoint. Text, image, video and audio models are available through one interface, and the API is OpenAI-compatible.

If one provider goes down, it automatically switches to another provider (Fallback), and it picks the more cost-effective option for each request. The service runs at the edge to reduce latency between users and inference. There are three integration options: calling the HTTP API directly, a typed Client SDK, and an Agent SDK for agents that handles tool use, loops and state.

It suits developers and teams who switch among multiple models or need to use several providers within one service. Fine-grained data policies can be set to restrict prompts so they go only to trusted models and providers.

What it is used for

Pricing and license

No subscription; buy credits and use them across any model or provider

Getting started

  1. Create an account at openrouter.ai (an organization for teams can be set up later).
  2. Buy credits.
  3. Create an API key.
  4. Send requests to the https://openrouter.ai/api/v1/chat/completions endpoint, or install an SDK with npm install @openrouter/sdk or pip install openrouter and make calls.
  5. Browse the model list at openrouter.ai/models and choose the slug of the model you will use.

Links

Related terms and companies

Recent mentions in briefings

Sources

Compiled from public sources (official pages, READMEs). Pricing and features change, so check the official site.