# OpenRouter

> A unified API for every major LLM. One endpoint, hundreds of models, with
> routing, fallbacks, and cost tracking across providers.

## What OpenRouter is and when to use it

OpenRouter is one OpenAI-compatible API in front of hundreds of models from
many providers. Send `POST https://openrouter.ai/api/v1/chat/completions`
with an OpenRouter API key and a `model` such as `openai/gpt-4o` or
`anthropic/claude-sonnet-4`. Every chat model accepts that same request shape.
Image, video, speech-to-text, and text-to-speech models use their own
endpoints, described in each model's per-model guide below.

Use OpenRouter when you need to:

- Call many models or providers through one API key, one bill, and one SDK.
- Switch models without changing code, or compare models on the same prompt.
- Route around provider outages and rate limits with automatic fallbacks.
- Use text, image, video, speech-to-text, and text-to-speech models from one place.

Getting started:

- API keys are self-serve: sign up at https://openrouter.ai and create a key at https://openrouter.ai/settings/keys. No sales contact is needed.
- Free tier: models with the `:free` variant cost nothing, and the free models router at `openrouter/free` picks one for you. Free usage is rate limited per day; buying credits raises the limit. See https://openrouter.ai/docs/api-reference/limits.md.
- Pricing is pay-as-you-go per token with no markup on the provider's price: https://openrouter.ai/pricing.
- MCP server for agents: https://mcp.openrouter.ai/mcp (OAuth in the browser, hosted by OpenRouter). Setup guide: https://openrouter.ai/docs/guides/overview/mcp-server.md.
- OpenAPI description of the whole API: https://openrouter.ai/openapi.json.

## Docs

- [Quickstart](https://openrouter.ai/docs/quickstart.md): Get an API key and send the first request.
- [API reference](https://openrouter.ai/docs/api-reference/overview.md): Endpoints, parameters, and response shapes.
- [Docs index](https://openrouter.ai/docs/llms.txt): Every documentation page as a markdown link with a one-line summary.
- [Full docs](https://openrouter.ai/docs/llms-full.txt): The complete documentation in one file, for agents that want everything at once.
- [Limits](https://openrouter.ai/docs/api-reference/limits.md): Credit limits, free-model request caps, and rate limits.
- [Free variant](https://openrouter.ai/docs/guides/routing/model-variants/free.md): Access free models with the `:free` variant.
- [Free models router](https://openrouter.ai/docs/guides/routing/routers/free-router.md): Route to whichever free model is available.
- [FAQ](https://openrouter.ai/docs/faq.md): Common questions about billing, privacy, routing, and models.

## Models

- [Models](https://openrouter.ai/models): Compare every model on OpenRouter by pricing, context length, and benchmarks, all through one API.
- [Model API](https://openrouter.ai/api/v1/models): Machine-readable list of every model, with pricing and supported parameters.
- Per-model guide: `https://openrouter.ai/{author}/{model}/llms.txt` returns one machine-readable guide with sections for every confirmed OpenRouter API that serves a public model: text/chat, image, video, speech-to-text, and text-to-speech.

## Site

- [Rankings](https://openrouter.ai/rankings): LLM leaderboards by real-world usage, ranked by tokens processed through the OpenRouter API, with per-modality views.
- [Pricing](https://openrouter.ai/pricing): How OpenRouter pricing works, with pay-as-you-go and enterprise plans, plus answers to common billing questions.
- [Providers](https://openrouter.ai/providers): The network of model providers available on OpenRouter, with a page per provider.
- [Compare](https://openrouter.ai/compare): Compare models by benchmarks, price, context length, latency, uptime, and throughput.
- [Collections](https://openrouter.ai/collections): Curated collections of models, including free models and image generation models.
- [Apps](https://openrouter.ai/apps): App and agent rankings, listing the apps built on OpenRouter and the top apps by category.
- [State of AI](https://openrouter.ai/state-of-ai): An empirical study of real-world LLM usage across more than 100 trillion tokens on OpenRouter.
- [Data](https://openrouter.ai/data): OpenRouter's empirical AI usage data for researchers, journalists, and academics.
- [Support](https://openrouter.ai/support): Frequently asked questions about using OpenRouter, and how to reach support.
- [For AI crawlers](https://openrouter.ai/for-ai): Text-only HTML index of the site with the top models per modality, every provider, and every collection, each with a description and canonical link.

## Ori

- [Ori Eval](https://openrouter.ai/ori/eval.md): Find the best model for your project — run your agent on your own prompts, on a harness and model that stay pinned, and rank the candidates.
- [spawn-ori-eval skill](https://openrouter.ai/skills/spawn-ori-eval): For coding agents. Read it to run a model eval by handing it to Ori, so the eval is authored and graded on a pinned harness and model rather than one you improvised.
- [install-ori-harness skill](https://openrouter.ai/skills/install-ori-harness): For coding agents. Read it to install Ori, sign in, run an existing agent CLI through Ori, choose an OpenRouter model, and verify the setup.
- [Install Ori](https://openrouter.ai/labs/ori/install.sh): `curl -fsSL https://openrouter.ai/labs/ori/install.sh | bash`
- [Ori Eval docs](https://openrouter.ai/docs/guides/ori/eval.md): Writing `*.eval.ts` files, LLM-as-judge grading, and model bakeoffs.

## Skills

- [Skills](https://github.com/OpenRouterTeam/skills): Every skill OpenRouter publishes. Each one is served as raw markdown from `https://openrouter.ai/skills/<name>` (also at `/.well-known/skills/<name>/SKILL.md`, for agents that look there first).
- [Agent guide](https://openrouter.ai/agents.md): Using OpenRouter from a coding agent.

## Benchmarks

- [Benchmarks](https://openrouter.ai/benchmarks): Browse live independent benchmark results for OpenRouter models, providers, and search engines.
- [GET /api/v1/benchmarks](https://openrouter.ai/api/v1/benchmarks): Fetch benchmark results programmatically with a valid OpenRouter API key.
- [Benchmarks API reference](https://openrouter.ai/docs/api/api-reference/benchmarks/list-benchmarks): Read the API contract for listing benchmark results and supported filters.