API Reference
Seedance 2.5 API: slugs and pricing
Seedance 2.5 in the Clipia API — T2V and I2V slugs, parameter limits, how to check availability in the catalog, estimate the price and submit over REST.
Seedance 2.5 is a video model from ByteDance. The API exposes it in two public modes: text-to-video and image-to-video.
| Mode | API slug |
|---|---|
| Text-to-video | seedance-2-5-t2v |
| Image-to-video | seedance-2-5-i2v |
Check availability in your environment
The model is enabled separately from code releases; both slugs are present in the production catalog (snapshot of November 4, 2026). Use a slug only when GET /v1/models returns it for your API key: a model that is switched off disappears from the list, and submitting a generation returns 404 with the code not_found.
To generate a video without the API, open the Seedance 2.5 landing: it appears in the Clipia catalog only when both modes are active.
Check availability
curl -s https://api.clipia.ai/v1/models \
-H "Authorization: Bearer $CLIPIA_KEY" \
| jq '.data[] | select((.slug // "") | startswith("seedance-2-5"))'LLM models in the same list have no slug, so the examples check for it — see Model catalog.
Parameter limits
A snapshot of input_schema as of November 4, 2026. GET /v1/models/{slug} returns the exact fields for the selected mode — treat that live schema as authoritative over this static summary.
| Parameter | Values |
|---|---|
prompt | up to 20,480 characters. The model advertises 30000, but does not execute longer prompts, so the API accepts 20,480 |
image_url | seedance-2-5-i2v only, required: the source image as an HTTPS URL |
duration | 4, 5, 6, 8, 10, 12, 15, 20, 25 or 30 seconds; default 4 |
resolution | 480p, 720p, 1080p; default 720p |
aspect_ratio | 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, adaptive; default 16:9 for T2V and adaptive for I2V |
reference_video_urls | up to 10 URLs; a request with more videos is rejected with 400 before any credits are reserved |
reference_image_urls, reference_audio_urls | model references: up to 30 images and up to 10 audio clips. The schema does not declare them; a request above the limit is not rejected, and the extra URLs are silently dropped |
generate_audio | sound is on by default; generate_audio: false returns a silent video. If the audio track does not pass the safety check, the video arrives without sound with an audio_dropped warning |
The defaults are execution defaults: a request without a parameter is priced and executed with exactly these values.
Fetch schema and price
curl https://api.clipia.ai/v1/models/seedance-2-5-t2v \
-H "Authorization: Bearer $CLIPIA_KEY"Before submitting, send the same input to POST /v1/models/seedance-2-5-t2v/estimate — see Cost estimation. The cost depends on duration and resolution; audio does not change it.
Pricing
Prices are current as of November 2026. The exact cost is always shown on the run button before anything is charged.
Submit text-to-video
curl -X POST https://api.clipia.ai/v1/models/seedance-2-5-t2v \
-H "Authorization: Bearer $CLIPIA_KEY" \
-H "Content-Type: application/json" \
-d '{
"input": {
"prompt": "One continuous cinematic tracking shot through a storm",
"duration": 10,
"resolution": "1080p",
"aspect_ratio": "16:9",
"generate_audio": true
}
}'For image-to-video, use seedance-2-5-i2v and pass the image in the image_url field, as declared by its input_schema. Don't copy field names from other integrations.
Explicit model swap
If Seedance 2.5 is missing from the live catalog, you can explicitly choose seedance-2-fast-t2v or seedance-2-fast-i2v. Don't swap the model silently when the user picked version 2.5.
Model catalog
GET /v1/models — Clipia image, video and audio models with exact slugs, input schema and pricing, including slugs that contain a slash.
Cost estimation
POST /v1/models/{model}/estimate — find out what a generation costs in credits before running it; no job is created and no credits are reserved.