Skip to content

Latest commit

 

History

2 Commits

Folders and files

Repository files navigation

Seedance 2.5 Flash API: cURL, Python, and Node.js Examples

Seedance 2.5 Flash on PoYo API docs Check

Runnable examples for ByteDance Seedance 2.5 Flash through PoYo's async video API: 4–30 second 480p or 720p videos with synchronized audio from text, first and last frames, or image, video, and audio references.

Answer: To use the Seedance 2.5 Flash API, send a server-side request to POST https://api.poyo.ai/api/generate/submit with model seedance-2.5-flash and a PoYo API key, then poll GET https://api.poyo.ai/api/generate/status/{task_id}. The examples in this repository run with Python 3.9+, Node.js 22+, or cURL.

Last verified against the live PoYo schema and pricing on November 16, 2026.

Seedance 2.5 Flash at a glance

  • Model ID: seedance-2.5-flash on POST https://api.poyo.ai/api/generate/submit
  • Released: November 12, 2026 by ByteDance; live on PoYo since November 16, 2026
  • Output: 480p or 720p (default 720p), 4–30 seconds, with synchronized audio on by default
  • Inputs: text, a first frame with an optional last frame, or up to 10 reference images, 5 reference videos, and 5 reference audio clips
  • Aspect ratios: auto, 21:9, 16:9, 4:3, 1:1, 3:4, 9:16
  • Price: from 18.5 credits ($0.0925) per second at 480p and 41 credits ($0.205) per second at 720p
  • Speed: nearly twice as fast as Seedance 2.5, according to ByteDance

Seedance 2.5 Flash API examples (video)

Real Seedance 2.5 Flash outputs generated through PoYo. The previews are silent GIFs; click one to open the MP4 with its generated audio.

Seedance 2.5 Flash API text-to-video example: paper boat in a rainy street at dusk
Text-to-video · 480p · 4 s · 74 credits
The output of payloads/seedance-2.5-flash-text-to-video.json
Seedance 2.5 Flash API example: Close-up to pull-back with rain ambience and violin
Text-to-video · 720p · 8 s · 328 credits
Close-up to pull-back with rain ambience and violin
Seedance 2.5 Flash API example: Five timestamped beats with dialogue, waves, and piano
30-second single take · 720p · 1,230 credits
Five timestamped beats with dialogue, waves, and piano
Seedance 2.5 Flash API example: Handheld shot with spoken line and kitchen sounds
Vertical 9:16 · 480p · 10 s · 185 credits
Handheld shot with spoken line and kitchen sounds
Prompt: paper boat (text-to-video, 480p, 4 s)
A paper boat drifts down a rain-filled street gutter at dusk, neon reflections rippling on the water, soft rain sounds.
Prompt: rainy violinist (720p, 8 s)
A street violinist in a dark wool coat plays under an old lamppost on a rainy cobblestone street at dusk. Close-up of fingers pressing the strings and the bow catching the warm light, then a slow pull-back reveals glowing shop windows, reflections on wet stones, and passersby with umbrellas slowing down to listen. Natural rain ambience and a gentle, expressive violin melody.
Prompt: 30-second timestamped story (720p, 30 s)
One continuous 30-second story of the same young woman in a yellow raincoat delivering a handwritten letter across a small seaside town at dawn. 0-6s: wide aerial shot of the sleepy harbor town, fishing boats rocking, soft morning light; she rides an old bicycle down a cobblestone hill. 6-12s: tracking shot beside her as she weaves past a fish market, seagulls lifting off, a vendor waves at her. 12-18s: low-angle shot as she climbs a steep stone staircase carrying the bicycle, breathing hard, the letter tucked in her pocket. 18-24s: she knocks on a blue door at the top; an elderly man opens it and she hands him the letter. 24-30s: close-up of the man reading, his eyes filling with tears; he smiles and says softly, "She remembered." Camera slowly pulls back to reveal the sunrise over the sea. Natural ambient sound: waves, gulls, bicycle bell, footsteps, gentle piano music. No subtitles, no on-screen text.
Prompt: vertical bakery (9:16, 480p, 10 s)
Vertical handheld shot: a young baker in a small morning bakery pulls a tray of golden croissants from the oven, steam rising in the warm light, then turns to the camera with a proud smile and says, "Fresh out of the oven." Gentle kitchen sounds and soft acoustic background music.

What is Seedance 2.5 Flash?

Seedance 2.5 Flash is the faster, lower-cost tier of ByteDance Seedance 2.5. According to ByteDance, it generates nearly twice as fast as Seedance 2.5. It keeps the Seedance 2.5 request shape at 480p and 720p, with synchronized dialogue, sound effects, and music generated in the same pass.

Seedance 2.5 Flash Seedance 2.5
Model ID seedance-2.5-flash seedance-2.5
Resolution 480p, 720p (default 720p) 480p, 720p, 1080p
Duration 4–30 seconds 4–30 seconds
Synchronized audio On by default Optional
Reference images / videos / audio up to 10 / 5 / 5 up to 30 / 10 / 10
480p, no reference video 18.5 credits ($0.0925) per second 28 credits ($0.14) per second
720p, no reference video 41 credits ($0.205) per second 63 credits ($0.315) per second

Choose Flash for drafts, high-volume social clips, and fast iteration. For 1080p delivery or more than 10 reference images, use Seedance 2.5.

Quickstart

  1. Create a key in the PoYo dashboard.
  2. Copy .env.example to .env and replace the placeholder.
  3. Run one of the examples. Each submits the task and polls until finished or failed.
POYO_API_KEY=YOUR_POYO_API_KEY_HERE
POYO_BASE_URL=https://api.poyo.ai

Python

python python/generate.py
python python/generate.py payloads/seedance-2.5-flash-reference-to-video.json

python/generate.py uses only the Python standard library.

Node.js

node node/index.mjs
node node/index.mjs payloads/seedance-2.5-flash-image-to-video.json

cURL

curl -X POST "https://api.poyo.ai/api/generate/submit" \
  -H "Authorization: Bearer $POYO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-flash",
  "input": {
    "prompt": "A paper boat drifts down a rain-filled street gutter at dusk, neon reflections rippling on the water, soft rain sounds.",
    "duration": 4,
    "resolution": "480p",
    "aspect_ratio": "16:9"
  }
}'

Status polling and the other modes are in curl/generate.md.

Example requests

Every file in payloads/ is a complete request body with public input media, so it runs as-is.

File Mode Settings Credits
seedance-2.5-flash-text-to-video.json Text-to-video 480p, 4 s, audio on 74 ($0.37)
seedance-2.5-flash-image-to-video.json First frame 480p, 4 s, aspect_ratio: auto 74 ($0.37)
seedance-2.5-flash-reference-to-video.json @Image1 + @Video1 + @Audio1 480p, 4 s + reference video 69 ($0.345)

Input modes

The supplied media selects the mode automatically.

Mode Selected when
Reference-to-video reference_image_urls, reference_video_urls, or reference_audio_urls is non-empty
Image-to-video (first / last frame) image_urls is non-empty
Text-to-video No media is provided

Send image_urls and the reference fields in separate requests. In first/last-frame mode, keep aspect_ratio at auto so the output follows the first frame, and use a last frame with the same shape. In the prompt, refer to reference media by upload order as @Image1, @Video1, and @Audio1, and give each item a role.

Writing Seedance 2.5 Flash prompts

  • Long single takes: split a 30-second prompt into timestamped beats, such as 0-6s: wide aerial shot of the harbor town ... 24-30s: close-up of the man reading, and keep one subject and setting across beats.
  • Audio in the same pass: describe dialogue in quotes, plus ambience, sound effects, and music. Add "no background music" to drop a layer, or set generate_audio to false for a silent clip.
  • References with roles: name every input by upload order and give it one job, for example "@Image1 is the cat, @Image2 is the scarf, match the rhythm of @Audio1".
  • First and last frames: describe the change between the two frames, keep aspect_ratio at auto, and use two images of the same shape.
  • Draft, then render: iterate at 480p, then render the final cut at 720p, or switch the model ID to seedance-2.5 for 1080p.

A complete 30-second timestamped prompt is in the API reference.

Request fields

Field Requirement Meaning
prompt required Video instructions; reference media as @Image1, @Video1, @Audio1.
duration required Integer from 4 to 30 seconds.
resolution optional 480p or 720p. Default 720p.
aspect_ratio optional auto, 21:9, 16:9, 4:3, 1:1, 3:4, or 9:16. Default auto.
generate_audio optional Synchronized audio. Default true; false returns video only.
image_urls optional First frame, or first and last frame. JPEG, PNG, or WebP up to 30 MB each.
reference_image_urls optional Up to 10 images. JPEG, PNG, or WebP up to 30 MB each.
reference_video_urls optional Up to 5 videos. MP4 or MOV, 2–30 s and up to 200 MB each; 30 s combined.
reference_audio_urls optional Up to 5 clips with reference images or videos. MP3 or WAV, 2–30 s and up to 15 MB each; 30 s combined.

Pricing

Billing is per second. 1 credit = $0.005.

Resolution No reference video With reference video
480p 18.5 credits ($0.0925) per second 11.5 credits ($0.0575) per billed second
720p 41 credits ($0.205) per second 25.5 credits ($0.1275) per billed second

Without reference video, billed seconds equal the output duration. With reference_video_urls, billed seconds are the output duration plus each reference video's duration floored to whole seconds. Confirm current prices on the model page before a production rollout.

Cost per clip

Without reference videos:

Clip length 480p 720p
5 seconds 92.5 credits ($0.4625) 205 credits ($1.025)
10 seconds 185 credits ($0.925) 410 credits ($2.05)
30 seconds 555 credits ($2.775) 1,230 credits ($6.15)

Failed generation tasks are not charged credits.

Async production flow

submit -> persist data.task_id -> poll during development -> receive callback_url webhook in production

Treat finished and failed as terminal states. Accept a webhook only for task IDs your system created, process it idempotently, and reconcile the final state through the status endpoint. Copy output videos to your own storage after the task finishes.

Common errors

Always read code in the response body: parameter errors return HTTP 200 with code: 400.

code error.type Example and fix
400 validation_error duration must be between 4 and 30 or resolution must be 480p or 720p: correct the field and resubmit.
400 content_moderation_error The prompt or media was declined by safety review: revise and submit a new task.
401 authentication_error Missing or invalid key: send Authorization: Bearer <POYO_API_KEY>.
402 insufficient_credits_error Top up credits in the dashboard.
429 rate_limit_error Back off and retry.

A task that fails after submission returns status: "failed" with error_message, and is not charged credits. See error codes.

FAQ

Is Seedance 2.5 Flash free to use? It is pay-as-you-go on PoYo, with no subscription or minimum spend. Eligible new accounts receive trial credits at sign-up, which you can use in the Playground before integrating the API.

How much does Seedance 2.5 Flash cost per second? 18.5 credits ($0.0925) per second at 480p and 41 credits ($0.205) per second at 720p. With reference videos, the rates are 11.5 and 25.5 credits per billed second, where billed seconds include the reference video length.

How is Seedance 2.5 Flash different from Seedance 2.5? Both take the same request fields and produce 4–30 second videos with audio. Flash outputs 480p or 720p and accepts up to 10 images, 5 videos, and 5 audio clips; Seedance 2.5 adds 1080p and accepts up to 30, 10, and 10.

Can Seedance 2.5 Flash make 1080p or 4K video? Flash outputs 480p or 720p, with 720p as the default. For native 1080p, switch the model ID to seedance-2.5, which takes the same request fields.

How long does a generation take? In PoYo testing on November 16, 2026, a 4-second 480p clip finished in about two minutes and a 30-second 720p clip in under three minutes, end to end. Times vary with load and settings.

Can I turn the audio off? Yes. Audio is generated by default; set generate_audio to false for a silent clip, or drop individual layers in the prompt.

How many reference files can one request use? Up to 10 reference images, 5 reference videos, and 5 reference audio clips. Reference videos together can last up to 30 seconds, and so can reference audio clips. Audio references are used together with at least one reference image or video.

Can I use it from an AI agent? Yes. The PoYo MCP bridge exposes the model as the poyo_seedance_2_5_flash tool.

Repository map

Related repositories

Links

License

MIT

Releases

Packages

Contributors

Languages