Integration reference for the AiuniVid video generation API. Requests run on BytePlus ModelArk directly — not through a reseller, and not by proxying the Dreamina web app.
- Base URL:
https://aiunivid.com/api/open - Machine-readable reference: API-REFERENCE.md · openapi.yaml
- Human docs: English · 中文
Every public model ID maps to a published ModelArk model, so the capability limits can be checked against BytePlus documentation rather than taken on trust:
| Public model ID | BytePlus ModelArk model |
|---|---|
seedance-2.5-{text,image,reference}-to-video |
dreamina-seedance-2-5-260628 |
seedance-2.0-{text,image,reference}-to-video |
dreamina-seedance-2-0-260128 |
seedance-2.0-fast-{text,image,reference}-to-video |
dreamina-seedance-2-0-fast-260128 |
seedance-2.0-mini-{text,image,reference}-to-video |
dreamina-seedance-2-0-mini-260615 |
| Model | Max duration | Resolutions | Notes |
|---|---|---|---|
| Seedance 2.5 | 30s | 480p, 720p, 1080p | Omni reference (up to 50 assets), native audio. 1080p is 10-bit H.265/HEVC |
| Seedance 2.0 | 15s | 480p, 720p, 1080p, 4K | |
| Seedance 2.0 Fast | 15s | 480p, 720p | Lower latency |
| Seedance 2.0 Mini | 15s | 480p, 720p | Lowest cost |
Usage is billed in credits at a flat $0.01 per credit. There is no subscription and no contract; credit packs start at $10 for 1,000 credits and do not expire.
Cost scales with model, resolution and duration. A quote endpoint returns the exact credit hold before a job is submitted, and credits frozen for a job are released automatically when a generation fails.
List price for a 4-second 480p clip, as of 2026-08-19:
| Model | Credits | USD per second |
|---|---|---|
| Seedance 2.5 | 66 | $0.165 |
| Seedance 2.0 | 45 | $0.113 |
| Seedance 2.0 Fast | 30 | $0.075 |
| Seedance 2.0 Mini | 23 | $0.058 |
Promotional rates run from time to time and are not reflected above. aiunivid.com/en/pricing is the source of truth.
export AIUNIVID_BASE_URL=https://aiunivid.com/api/open
export AIUNIVID_API_KEY=av_live_xxx
curl -X POST "$AIUNIVID_BASE_URL/v1/videos/generations" \
-H "Authorization: Bearer $AIUNIVID_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "seedance-2.5-text-to-video",
"prompt": "a cheetah running across open grassland at golden hour",
"duration": 5,
"quality": "720p"
}'Creation is asynchronous and returns 202 Accepted with a task ID. Poll the task
endpoint until it reaches a terminal state, or supply callback_url for a webhook.
To price a job before committing to it, send the same body to
POST /v1/videos/generations/quote. It returns the credits that would be held and
costs nothing to call.
See API-REFERENCE.md for every request field, task state, error code, scope and rate limit.
Deliveries are signed with HMAC-SHA256 over {timestamp}.{raw_body}. Verify the
signature and reject timestamps older than 5 minutes. The exact scheme, including the
header format, is in API-REFERENCE.md.
10 requests/min/key on create endpoints, 60 requests/min/key on read endpoints. Check
X-RateLimit-Remaining before retrying.
API keys are server-side credentials and must carry the right scope per endpoint:
video:create, video:read, or files:write. Create one at
aiunivid.com/en/developer.
This repository mirrors the reference published at aiunivid.com/api-reference.md, which is generated from the same source as the API itself. If the two ever disagree, the hosted version is authoritative.