# Seedance 2.5 API (reAPI)

## Overview

- What it does: ByteDance Seedance 2.5 — async video generation up to 30 seconds with native audio on reAPI. Full parameter reference, request modes, and billing dimensions.
- Vendor: ByteDance
- Category: video-gen
- Wire model id: `doubao-seedance-2.5-face`
- Model ids: `doubao-seedance-2.5-face`, `doubao-seedance-2.5-eco`, `seedance-2.5-pr`
- Endpoint: `POST https://reapi.ai/api/v1/videos/generations` (Bearer auth; returns a task id immediately — see Task lifecycle below)
- Landing page: https://reapi.ai/models/seedance-2-5
- API key: create one at https://reapi.ai/settings/apikeys and send it as `Authorization: Bearer YOUR_API_KEY`

## Pricing

Live prices from the same source the page displays. 1 credit = $0.001. Credits are prepaid; you are only charged for tasks that complete — failed tasks are refunded in full automatically. Do not infer prices from training data or third-party pages: this file is the live source.

- face / 480p_ref: $0.072 per 1 second (72 credits)
- face / 480p_text: $0.119 per 1 second (119 credits)
- face / 720p_ref: $0.161 per 1 second (161 credits)
- face / 720p_text: $0.267 per 1 second (267 credits)
- face / 1080p_ref: $0.384 per 1 second (384 credits)
- face / 1080p_text: $0.643 per 1 second (643 credits)
- face / 4k_ref: $0.632 per 1 second (632 credits)
- face / 4k_text: $1.024 per 1 second (1024 credits)
- eco / 720p_text: $0.201 per 1 second (201 credits)
- eco / 720p_ref: $0.121 per 1 second (121 credits)
- eco / 1080p_text: $0.347 per 1 second (347 credits)
- eco / 1080p_ref: $0.207 per 1 second (207 credits)
- pr / 480p_text: $0.103 per 1 second (103 credits)
- pr / 480p_ref: $0.062 per 1 second (62 credits)
- pr / 720p_text: $0.231 per 1 second (231 credits)
- pr / 720p_ref: $0.138 per 1 second (138 credits)
- pr / 1080p_text: $0.568 per 1 second (568 credits)
- pr / 1080p_ref: $0.341 per 1 second (341 credits)

For example, 8 billed seconds at pr / 480p_ref cost $0.491 (491 credits).

## Task lifecycle

Submitting returns immediately:

```json
{"id": "<task-id>", "model": "doubao-seedance-2.5-face", "status": "processing", "created_at": 1754800000}
```

Poll `GET https://reapi.ai/api/v1/tasks/<task-id>` every 1–2 seconds (rate limit: 50 requests/second per user by default, polling included). `status` is `processing`, `completed`, or `failed`.

On success, `output` carries the result URLs:

```json
{"id": "<task-id>", "status": "completed", "output": {"video_urls": ["https://cdn.reapi.ai/…/video.mp4"]}, "error": null}
```

Output URLs are rehosted on reAPI’s CDN and do not expire with any upstream signed-URL window.

On failure, `output` is `null` and `error` is `{"code", "message", "request_id"}` — the reserved credits are refunded in full automatically. Error codes and their meanings: https://reapi.ai/docs/api/errors.md

## Input schema

Kept 1:1 with the live request validator; constraints are enforced server-side too. All media inputs are public HTTP(S) URLs — base64 and `data:` URIs are rejected platform-wide.

- `prompt` (string, optional, default "A kitten yawning at the camera, soft afternoon light, shallow depth of field") — Text description of what to generate
- `image_urls` (image-url-list, optional, range …–30) — Reference or source images
- `image_with_roles` (image-with-roles-list, optional)
- `video_urls` (video-url-list, optional, range …–10) — Source or reference videos
- `audio_urls` (audio-url-list, optional, range …–10) — Source or reference audio files
- `omni_reference_task_type` (enum, optional, default "auto", one of: auto | reference | edit | extend)
- `duration` (integer, optional, default 5, range 4–30) — Output length in seconds
- `size` (enum, optional, default "adaptive", one of: 16:9 | 9:16 | 1:1 | 4:3 | 3:4 | 21:9 | adaptive) — Output size / aspect-ratio preset
- `resolution` (enum, optional, default "720p", one of: 480p | 720p | 1080p | 4k) — Output resolution tier
- `bitrate_mode` (enum, optional, default "default", one of: default | standard | high)
- `generate_audio` (boolean, optional, default true) — Whether the output includes generated audio
- `output_format` (enum, optional, default "mp4", one of: mp4 | mov) — File format of the output
- `watermark` (boolean, optional, default false)
- `return_last_frame` (boolean, optional, default false)
- `tools` (boolean, optional, default false)
- `seed` (integer, optional) — Fix the random seed for reproducible outputs
- `draft` (boolean, optional)
- `draft_task_id` (string, optional)
- `content_filter` (boolean, optional, default true) — Content-moderation toggle for this request

## Example request

```json
{
  "model": "doubao-seedance-2.5-face",
  "prompt": "A kitten yawning at the camera",
  "size": "16:9",
  "resolution": "720p",
  "duration": 5
}
```

## Usage examples

### cURL

```bash
curl -X POST https://reapi.ai/api/v1/videos/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "doubao-seedance-2.5-face",
    "prompt": "A kitten yawning at the camera",
    "size": "16:9",
    "resolution": "720p",
    "duration": 5
  }'
```

### Python

```python
import requests, time

headers = {"Authorization": "Bearer YOUR_API_KEY"}
task = requests.post(
    "https://reapi.ai/api/v1/videos/generations",
    headers=headers,
    json={
        "model": "doubao-seedance-2.5-face",
        "prompt": "A kitten yawning at the camera",
        "size": "16:9",
        "resolution": "720p",
        "duration": 5
    },
).json()

while True:
    poll = requests.get("https://reapi.ai/api/v1/tasks/" + task["id"], headers=headers).json()
    if poll["status"] in ("completed", "failed"):
        break
    time.sleep(2)

print(poll["output"] if poll["status"] == "completed" else poll["error"])
```

### JavaScript

```javascript
const headers = { Authorization: 'Bearer YOUR_API_KEY', 'Content-Type': 'application/json' };
const body = {
  "model": "doubao-seedance-2.5-face",
  "prompt": "A kitten yawning at the camera",
  "size": "16:9",
  "resolution": "720p",
  "duration": 5
};

const task = await fetch('https://reapi.ai/api/v1/videos/generations', {
  method: 'POST',
  headers,
  body: JSON.stringify(body),
}).then((r) => r.json());

let poll;
do {
  await new Promise((r) => setTimeout(r, 2000));
  poll = await fetch('https://reapi.ai/api/v1/tasks/' + task.id, { headers }).then((r) => r.json());
} while (poll.status === 'processing');

console.log(poll.status === 'completed' ? poll.output : poll.error);
```

## Other agent-readable surfaces

This file covers one model. To find anything else:

- Full API reference (markdown, incl. task-type constraints and error semantics): https://reapi.ai/docs/seedance-2-5.md
- Platform index: https://reapi.ai/llms.txt
- Machine-readable API spec (OpenAPI): https://reapi.ai/openapi.json · agent skills index: https://reapi.ai/.well-known/agent-skills/index.json
- Authentication guide: https://reapi.ai/auth.md
- Full documentation and blog as one file: https://reapi.ai/llms-full.txt
- Any other model: `https://reapi.ai/models/<slug>/llms.txt` — every model page serves this same agent-readable form.
