> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.empiriolabs.ai/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.empiriolabs.ai/_mcp/server.

# MOSS Video and Audio

> Open-source 32B MoE foundation model that generates synchronized video and audio in one inference step with precise dual-tower lip-sync.

![MOSS Video and Audio](https://media.empiriolabs.ai/model-logos/moss-video-and-audio.png)

[OpenMOSS](/providers/openmoss) · Video Generation

`POST /v1/videos/generations`

Open-source 32B MoE foundation model that generates synchronized video and audio in one inference step with precise dual-tower lip-sync.

## At a glance

| Field               | Value                                      |
| ------------------- | ------------------------------------------ |
| Model id            | `moss-video-and-audio`                     |
| Model release date  | 2026-01-29                                 |
| Input modalities    | Text, Image                                |
| Output modalities   | Video, Audio                               |
| Context window      | -                                          |
| Weight precision    | BF16                                       |
| Features            | audio_sync, lipsync                        |
| Native inference    | Yes                                        |
| New                 | No                                         |
| Supported endpoints | `POST /v1/videos/generations`              |
| Alternate model ids | `moss-video-audio`, `openmoss/video-audio` |

## Pricing

| Charge      | Spec           | Rate    |
| ----------- | -------------- | ------- |
| 360p Video  | per second     | \$0.12  |
| 720p Video  | per second     | \$2.50  |
| T2V Fast    | additional fee | \$0.065 |
| T2V Quality | additional fee | \$0.13  |

## Example request

```bash
curl https://api.empiriolabs.ai/v1/videos/generations \
  -H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{"model": "moss-video-and-audio", "prompt": "sunrise over the ocean", "duration": 6}'
```

## Parameters

| Parameter             | Type    | Required | Default       | Description                                                                                                                                  |
| --------------------- | ------- | -------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------- |
| `prompt`              | string  | yes      | -             | Scene description. With image attached, becomes an image-to-video prompt.                                                                    |
| `mode`                | enum    | no       | `"auto"`      | auto runs i2v when an image is attached, otherwise t2v. t2v: text-to-video. i2v: animate the attached image. · Allowed: `auto`, `t2v`, `i2v` |
| `resolution`          | enum    | no       | `"360p"`      | Output resolution. · Allowed: `360p`, `720p`                                                                                                 |
| `aspect_ratio`        | enum    | no       | `"landscape"` | Text-to-video only: landscape (16:9) or portrait (9:16). Image-to-video follows the image. · Allowed: `landscape`, `portrait`                |
| `duration`            | integer | no       | `4`           | Clip length in seconds, from 2 to 8. · Range: 2 – 8                                                                                          |
| `t2v_quality`         | enum    | no       | `"quality"`   | Text-to-video only. fast trades fidelity for \~2× speed. · Allowed: `fast`, `quality`                                                        |
| `num_inference_steps` | integer | no       | `25`          | Diffusion steps. More = higher fidelity, slower. · Range: 10 – 50                                                                            |
| `cfg_scale`           | number  | no       | `5.0`         | Classifier-free guidance. Higher = follows prompt more strictly. · Range: 1.0 – 10.0                                                         |
| `sigma_shift`         | number  | no       | `5.0`         | Schedule shift. Only valid when resolution=360p. · Range: 1.0 – 10.0                                                                         |
| `image`               | string  | no       | -             | Reference image URL for i2v mode.                                                                                                            |
| `negative_prompt`     | string  | no       | `""`          | What to avoid.                                                                                                                               |
| `seed`                | number  | no       | -             | Reproducibility seed. · Range: 0 – 4294967295                                                                                                |

## Notes

32B-parameter MoE with synchronized lip-sync video + audio in a single inference.

**Constraints**

* Generation can take 20+ minutes
* Image-to-Video typically yields superior results to text-to-video
* Only 1 image supported (used as the first frame)
* Video inputs NOT supported

**Image formats**

* jpg, jpeg, png, webp, heic, heif, bmp, tiff, tif, gif

---

*Machine-readable schema:* `GET https://api.empiriolabs.ai/v1/models/moss-video-and-audio`.