Dreamina Seedance 2.0 (International)

Models / Tiers

Per-second billing (4-15s), charged only on success and refunded on failure. Native resolution — no upscaling.

Price cut on 8 August 2026: Mini down ~56%, Fast down ~24%, standard tier unchanged. A job carrying a reference video bills on a different basis — a lower per-second rate, but charged on (source seconds + output seconds), so editing or continuing a clip into an output of the same length is not cheaper than generating it outright. Mini $0.0138 / $0.030, Fast $0.0378 / $0.0812 (480p / 720p).

Video Generation

# Text-to-video (model: dreamina-seedance-2-0 / -fast / -mini)
curl -X POST https://api.apimodels.app/v1/video/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "dreamina-seedance-2-0-fast",
    "prompt": "a hummingbird hovering near a bright flower, slow motion",
    "resolution": "720p",
    "aspect_ratio": "16:9",
    "duration": 5
  }'

# Real-person: register the face in the asset library first (account "intl", see below),
# then reference it — a RAW real face URL is rejected at create-time
# (InputImageSensitiveContentDetected.PrivacyInformation), asset:// passes:
#   "reference_image_urls": ["asset://asset-2026...-xxxx"]
#
# Image-to-video of a specific real person: pass the asset:// AS the first frame.
# (first_frame_url and reference_image_urls are mutually exclusive — when a first
#  frame is present, reference_image_urls is ignored, so put the person there):
#   "first_frame_url": "asset://asset-2026...-xxxx"

Asset Library (real person, asset://)

A photorealistic real face passed as a raw input image is rejected by upstream moderation ("may contain real person"). To make real-person video, register the face in the asset library to get an asset:// reference and pass it in. The group MUST be created with account: "intl" (same account/project, billing19, as intl generation) for the asset to resolve; each upload costs $0.01.

1. Create a group: POST /v1/assets/groups with account:"intl" in the body → get a group_id.
2. Register the face: POST /v1/assets (url + asset_type:"Image" + group_id) → get asset://… (the call waits until Active; failure returns HTTP 422 and is not charged).
3. Generate: put the asset:// in reference_image_urls (person as a reference, works with text-to-video) OR in first_frame_url (animate that portrait as the first frame — the two are mutually exclusive).

# 1) Create a group ON THE INTL account (account: "intl")
curl -X POST https://api.apimodels.app/v1/assets/groups \
  -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" \
  -d '{"name": "my-people", "account": "intl"}'
# → { "data": { "id": 1, "account": "intl", ... } }

# 2) Register a face from a public URL. The call WAITS until the asset is confirmed
#    usable (~3s), so the returned id is ready to use right away. Charged $0.01 ONLY
#    on success — if the image can't be fetched or fails moderation you get HTTP 422
#    and are NOT charged.
curl -X POST https://api.apimodels.app/v1/assets \
  -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" \
  -d '{"url": "https://example.com/face.jpg", "asset_type": "Image", "group_id": 1}'
# → { "data": { "id": "asset-2026...-xxxx", "asset_url": "asset://asset-2026...-xxxx", "status": "Active" } }
#   (on failure: HTTP 422, not charged)

# 3) Use asset_url in generation — either reference_image_urls (multi-image) or, to
#    animate that exact person, first_frame_url. No polling needed once Active.

Real-person video in the Playground

The Playground has two uploads: "Portrait" and "Reference images". Real people go through "Portrait" only — it auto-registers the face to asset:// and clears moderation, the one correct entry for real people. Four steps:

1. Pick a mode: Text-to-video, Multi-image reference, or Multimodal reference. Do NOT pick First-frame / First+last-frame — those ignore all references, including the portrait.
2. Upload one clear face photo in "Portrait" → auto-registered to asset://, moderation-safe. This single upload is all you need.
3. "Reference images" is optional and for non-real-person scene / style / object references only. Do not put a real face here — it is rejected by moderation (intl auto-retries by registering it to asset://, but the Portrait uploader is the clean path).
4. Write a prompt describing the action and shot → Generate.

Notes

• Real people always go through asset://. Never pass a raw real-person image in first_frame_url / reference_image_urls — a raw real face is rejected by upstream at create-time.
• First-frame and reference modes are mutually exclusive: when first_frame_url is present, reference_image_urls (including asset:// portraits) is ignored. To animate a specific person, put that portrait's asset:// in first_frame_url — do not pass a raw first frame plus an asset:// reference.
• account:"intl" assets are usable only by dreamina-seedance-2-0*; they are not interchangeable with the China official / sd-A assets.
• Async: create returns a taskId; poll GET /v1/video/generations?task_id= until completed. Result files are kept 7 days — download or re-host promptly.
• Use real-person material only with subjects you have the rights/consent to depict; portraits and outputs are screened by upstream moderation.

Try it in the Playground