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

# Media and embeddings

> Generate or transform image, audio, video, and embedding outputs through native paid inference.

Media routes use the same native payment behavior as text inference. Status polling routes are not billable; the charge happens on generation submission.

## Routes

| Method | Path                       | Description                               |
| ------ | -------------------------- | ----------------------------------------- |
| `POST` | `/v1/images/generations`   | Generate images from a prompt.            |
| `POST` | `/v1/images/edits`         | Edit an image with a prompt.              |
| `POST` | `/v1/audio/speech`         | Generate speech audio bytes.              |
| `POST` | `/v1/audio/transcriptions` | Transcribe base64 audio.                  |
| `POST` | `/v1/embeddings`           | Generate embeddings.                      |
| `POST` | `/v1/videos/generations`   | Submit video generation.                  |
| `GET`  | `/v1/videos/{id}`          | Poll video generation status.             |
| `GET`  | `/v1/videos/{id}/stream`   | Stream video status until terminal state. |

## Common headers

<ParamField header="Authorization" type="string">
  `Bearer compose-...` for Compose Key billing.
</ParamField>

<ParamField header="PAYMENT-SIGNATURE" type="string">
  Raw x402 payment signature for x402-aware clients.
</ParamField>

## Common fields

<ParamField body="model" type="string" required>
  Public catalog model ID.
</ParamField>

<ParamField body="prompt" type="string">
  Required for image and video generation.
</ParamField>

<ParamField body="input" type="string | array">
  Required for speech and embeddings.
</ParamField>

<ParamField body="file" type="string">
  Required for audio transcription. The route expects base64 content.
</ParamField>

<ParamField body="image" type="string">
  Required for image edits; also accepted as an initial image for some video models.
</ParamField>

<ParamField body="image_url" type="string">
  URL form of an input image.
</ParamField>

## Video status stream

<ParamField query="pollIntervalMs" type="number">
  Poll interval for `/v1/videos/{id}/stream`. Default `2000`, minimum `500`, maximum `30000`.
</ParamField>

<ParamField query="timeoutMs" type="number">
  Stream timeout. Default `600000`, minimum `5000`, maximum `3600000`.
</ParamField>

The stream emits `compose.video.status`, `compose.error`, and `data: [DONE]`.