> ## Documentation Index
> Fetch the complete documentation index at: https://mulerouter.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Welcome

> One API for every multimodal AI provider

**MuleRouter** is an all-in-one API service for developers. One API key
gives you access to multimodal AI from many providers — chat, image,
video, speech, music — through a single, unified interface. We negotiate
higher-tier service levels with each provider through aggregated
purchasing, so the same API tends to be more stable and faster than
going direct.

This site is the reference for the MuleRouter HTTP API, the official CLI,
and the rest of the platform surface.

## Where to start

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/docs/get-started/quickstart">
    Send your first request in under a minute — `curl`, an API key, done.
  </Card>

  <Card title="Authentication" icon="key" href="/docs/get-started/authentication">
    Get an API key, secure it, and use it in production.
  </Card>

  <Card title="Asynchronous tasks" icon="hourglass-half" href="/docs/get-started/async-tasks">
    The async pattern shared by every long-running endpoint.
  </Card>

  <Card title="Webhooks" icon="bell" href="/docs/get-started/webhooks">
    Receive signed callbacks instead of polling for status.
  </Card>

  <Card title="Error codes" icon="circle-exclamation" href="/docs/get-started/error-codes">
    HTTP status codes and MuleRouter error codes in one place.
  </Card>

  <Card title="API Reference" icon="code" href="/docs/api-reference/introduction">
    The full per-endpoint catalog — LLM, image, video, speech, music.
  </Card>

  <Card title="CLI" icon="terminal" href="/docs/cli/quickstart">
    `mulerouter` — drive every endpoint from the terminal.
  </Card>

  <Card title="Console" icon="user" href="https://www.mulerouter.ai/app">
    Manage API keys, billing, and webhooks in the web console.
  </Card>
</CardGroup>

## What you can build with MuleRouter

| Capability                 | Highlights                                                        | Where to look                                                                 |
| -------------------------- | ----------------------------------------------------------------- | ----------------------------------------------------------------------------- |
| **Chat / LLM**             | OpenAI-compatible Chat Completions across multiple model families | [`/v1/chat/completions`](/docs/api-reference/endpoint/openai/chat)                 |
| **Image generation**       | OpenAI, Google, Alibaba, Midjourney                               | [Image Generation](/docs/api-reference/endpoint/openai/gpt-image-2/generation)     |
| **Image editing**          | GPT Image 2, Nano Banana 2/Pro, Qwen Image Edit                   | [Image Editing](/docs/api-reference/endpoint/openai/gpt-image-2/edit)              |
| **Video generation**       | Wan 2.6 / 2.5 / 2.2 / 2.1, Kling V3, Midjourney Video             | [Video Generation](/docs/api-reference/endpoint/alibaba/wan2.6-t2v/generation)     |
| **Image / video to video** | Keyframe interpolation, VACE outpainting & inpainting             | [Wan 2.1 VACE](/docs/api-reference/endpoint/alibaba/wan2.1-vace-plus/generation)   |
| **Speech (TTS)**           | MiniMax Speech 2.8 HD / Turbo, 37+ languages                      | [Speech 2.8 HD](/docs/api-reference/endpoint/minimax/speech-2.8-hd/text-to-speech) |
| **Music**                  | MiniMax Music 2.5 / 2.0 with auto-lyrics                          | [Music 2.5](/docs/api-reference/endpoint/minimax/music-2.5/text-to-music)          |

## How the API is shaped

* **Providers under `/vendors`** — `/vendors/openai/...`, `/vendors/google/...`, etc. The path mirrors the upstream provider's own API surface as closely as possible, so most existing SDKs work with a base-URL swap. See [Endpoint conventions](/docs/get-started/endpoints).
* **Industry aliases** — `/v1/chat/completions` and `/v1/models` are aliases for the corresponding `/vendors/...` paths so OpenAI-flavored SDKs work out of the box.
* **Asynchronous tasks** — image and video generation use a `POST` → poll `GET` pattern. See [Asynchronous tasks](/docs/get-started/async-tasks).
* **Webhooks** — register an HTTPS endpoint to skip polling entirely. See [Webhooks](/docs/get-started/webhooks).

## Help

* Found a bug? Include the `Traceparent` response header when filing
  the issue — it lets us trace the exact request through our pipeline.
* Discord — community channel linked in the sidebar.
* Email — [contact@mulerouter.ai](mailto:contact@mulerouter.ai)