# Seedance 2.5 Draft to Final - Core.Today AI API

> Turn a Seedance 2.5 draft into the final 1080p video. Make a cheap 480p draft with bytedance/seedance-2.5 and draft: true, check it, then send that job's job_id here. The final reuses the draft's prompt, reference media, duration, aspect ratio, seed and audio setting, so what you approved is what you get.

- **Provider**: ByteDance
- **Model ID**: bytedance/seedance-2.5/draft-to-final
- **Category**: Video Generation
- **Credits**: 6680 per 5s draft — billed as 1080p for the draft's duration (1335 credits/s); drafts with reference videos: 799 credits per (input+output) second
- **Speed**: Medium
- **Quality**: Ultra

## Features
- Final video at 1080p (10-bit color, H.265/HEVC)
- Reuses the draft's prompt, reference media, duration, aspect ratio, seed and audio — no drift between preview and final
- Only output format and watermark can be set again
- Served directly from BytePlus ModelArk; never falls back to another provider

## Use Cases
- Iterating on composition and motion at 480p cost before paying for 1080p
- Client review workflows: share the draft, then render the approved version
- Batch exploration of prompts where only the winners are rendered in high quality

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **draft_job_id**: string - job_id of a completed Seedance 2.5 draft (POST /predictions with model bytedance/seedance-2.5 and draft: true), created by your team within the last 7 days. Not the upstream task id.

### Optional
- **output_format**: string (default: mp4) - Output video container format. Not reused from the draft — omitted means mp4.
  Options: mp4, mov
- **watermark**: boolean (default: false) - Add a watermark. Not reused from the draft — omitted means no watermark.

## Examples

### Finalize an approved draft
Step 2 of draft mode. Step 1 was POST /predictions with model bytedance/seedance-2.5, your prompt and inputs, and draft: true; its job_id goes here.
```json
{
  "model": "bytedance/seedance-2.5/draft-to-final",
  "input": {
    "draft_job_id": "YourDraftJobId0123456789abcdefAB",
    "output_format": "mp4"
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Do not send prompt, images, videos, audio, duration, aspect_ratio, seed or generate_audio — they come from the draft. To change any of them, make a new draft.
- The final is billed on the draft's settings: a 10-second draft finalizes as a 10-second 1080p video.
- 1080p output uses H.265/HEVC. If a player cannot open it, try VLC, mpv or QuickTime Player.
- A draft that ran on a fallback provider cannot be finalized — drafts never fall back, so this only affects requests made before draft mode existed.

## Documentation
https://docs.byteplus.com/en/docs/modelark/seedance-2-5#2.5_draft_mode