Become a founding member.

Paid

Seedance 2.5 Flash AI video generator

Copy agent promptDownload schemaOpen in a workflowCopy link

Seedance 2.5 Flash is an AI video generator from ByteDance. The cheaper tier of the Seedance 2.5 card (set Model to Seedance 2.5 Flash). Set it up below, then run it in a workflow or through the API.

Video~90 cr for 5s at default settings, up to ~540 cr for 30sByteDance

One 5s 720p clip is about 90 credits.

Input

16 fields

Result Idle

Your result will appear here.

Set your inputs, then open a workflow to create it.

Credit estimate uses the model defaults. Your workflow shows the final estimate for your inputs before you run.

Model guide

How to use Seedance 2.5 Flash

  1. Write a prompt that describes the shot.
  2. Add images, a video or audio if you want to guide the result. This is optional.
  3. Pick the length (4 to 30 seconds), the resolution (480p or 720p) and the aspect ratio (21:9, 16:9, 4:3, 1:1, 3:4 or 9:16).
  4. Choose Open in a workflow. Sign in, check the credit estimate on the board, then run it.
  5. Connect the result to the next step on the board, such as an editor or another model.

Inputs explained

Prompt (required)
Set the prompt used by this model. Type: text.
Start Frame (optional)
The video opens exactly on this image. Use it when you have already composed the shot (framing, background, camera position) and want it kept. Costs one of the registered materials (up to 8 per run, 5 on the older inline route). Pair it with an End Frame for a start-to-end shot. It cannot be combined with any reference media (Reference Images, Reference Audio, Audio): Seedance runs a start/end-frame task OR a reference task, never both (the mix is refused before you are charged). With a Reference Video wired, this image is sent as one more reference image instead, so the video no longer opens on it. A Start Frame shot cannot take your own voice: with Generate Audio on, the model makes its sound. Type: image.
Reference Image 1 (optional)
An identity and style guide, not a framing guide. The model may reframe freely, so a carefully composed shot wired here comes back as a DIFFERENT shot. If you want the video to open on this exact image, use Start Frame instead. Type: image.
Reference Video (optional)
Set the reference video used by this model. Type: list of video.
Reference Audio (optional)
Audio the model listens to; this track never plays in the result. In an edit (a Reference Video wired and a prompt that changes it, like "replace the man in @Video1"), wire the audio of that same clip here, turn Generate Audio off, and add this line to the prompt: "Match the framing of @Video1 in every frame: same face size and position, same distance from the camera. Never move closer or crop tighter." Without it, the audio can pull a wide shot into a close-up. The mouth then follows the timing and big shapes of this track, close but not frame perfect, and the clip comes back silent. Check its length against the clip (short clips have come back about 6 percent fast), retime it in the Video Editor if it is off, then lay the audio back in. Keep Generate Audio off when the track is a copyrighted song, because with it on Seedance can refuse the generated audio for copyright. With Generate Audio on, Seedance re-voices the track (timbre, timing and words), so the lips will not match it exactly. Without a Reference Video, it cannot be combined with a Start Frame (the mix is refused before you are charged). mp3 or wav only. Type: list of audio.
Aspect Ratio (optional)
Set the aspect ratio used by this model. Type: text. Options: 21:9, 16:9, 4:3, 1:1, 3:4, 9:16, adaptive. Default: "16:9".
Duration (optional)
Set the duration used by this model. Type: text. Options: Auto (model decides), 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30. Default: "5".
Resolution (optional)
Seedance 2.5 Flash makes 480p or 720p only. For 1080p or 4K, set Model back to Seedance 2.5. Type: text. Options: 480p, 720p. Default: "720p".
Generate Audio (optional)
On: the clip comes back with sound the model generates, guided by any Reference Audio. Off: the clip comes back SILENT, and it never keeps your own track. For lip sync in an edit, turn this off, wire the audio of the clip being edited into Reference Audio, and add this line to the prompt: "Match the framing of @Video1 in every frame: same face size and position, same distance from the camera. Never move closer or crop tighter." The mouth then follows the timing and big shapes of that track, close but not frame perfect; check the take length against the clip and retime it in the Video Editor if it is off, then lay the audio back in. Keep it off when that audio is a copyrighted song, because with it on Seedance can refuse the generated audio for copyright. Outside an edit, with this on, Seedance re-voices any Reference Audio, so the lips will not match your own track exactly. Type: boolean. Default: true.
Watermark (optional)
Set the watermark used by this model. Type: boolean. Default: false.
Content filter (optional)
On by default. Seedance screens your prompt, your reference material and the finished clip, so a run can fail after it has been submitted when something gets flagged (copyright is the common one). Turning this off asks Seedance to skip that safety and IP moderation. It costs 10 percent more per run, and you are responsible for holding the rights to whatever you generate. Type: boolean. Default: true.
Reference Image (from Seedance 2.0) (optional)
An image arriving from the Seedance 2.0 card. It is used as a reference, not as an opening frame: the model may reframe freely. To make the video open on an exact image, wire the Start Frame port above. Type: image.
End Frame (optional)
Optional last frame. Beside a Start Frame the video ends exactly on this image (and, like the Start Frame, it cannot be combined with reference media). With no Start Frame it is treated as one more reference image, so it guides the ending rather than pinning it. Type: image.
Reference Images (optional)
Identity and style guides. The model may reframe freely, so these steer WHO and WHAT is in the shot, not how it is framed; to pin the framing use Start Frame. Start Frame, End Frame and any reference VIDEO all draw on the same budget of registered materials per run (up to 8, or 5 on the older inline route). Type: list of image.
Audio (optional)
Optional reference audio. Shares the reference-audio slot with Reference Audio, and wiring both concatenates them: up to 2 tracks and 30 seconds between the two. It behaves exactly like Reference Audio (see that port for lip sync in an edit) and never plays in the result. Without a Reference Video, it cannot be combined with a Start Frame. Type: list of audio.
Aspect Ratio override (set by Seedance 2.0) (optional)
Leave this empty. It is filled in automatically when the Seedance 2.0 node routes a run into this model, and while it has a value it OVERRIDES the Aspect Ratio field above ("auto" lets the model choose from your reference media). To change the shape of your video, use Aspect Ratio, not this. Type: text. Options: 21:9, 16:9, 4:3, 1:1, 3:4, 9:16, auto.

Tips before you run

Start with the default settings and change one input at a time. Keep the result you like on your board so the next step can use it.

If you add a reference file, use a link the workflow can read. Check the inputs and credit estimate on the board before starting a run.

Seedance 2.5 Flash pricing

Seedance 2.5 Flash costs about 90 credits per 5s, 720p clip at default settings, and up to about 540 credits for 30 seconds. You pay with Wireflow credits, and the board shows the estimate for your settings before you run.

See Wireflow plans and credits

Seedance 2.5 Flash vs Seedance 2.5

DetailSeedance 2.5 FlashSeedance 2.5
MakerByteDanceByteDance
Price at default settings~90 credits per 5s, 720p clip~135 credits per 5s, 720p clip
Clip length4 to 30 seconds4 to 30 seconds
Resolution480p or 720p480p, 720p, 1080p or 4K
Works fromA prompt, images, a video or audioA prompt, images, a video or audio
SoundYes, on by defaultYes, on by default

Both run on one Wireflow board, so you can try the same prompt on each. Try Seedance 2.5.

Seedance 2.5 Flash API

Run Seedance 2.5 Flash from your own code with the Wireflow API. Open it in a workflow, then call the run endpoint with that workflow's ID and poll for the result. Agents can run it through MCP.

# Model: Seedance 2.5 Flash. Open a workflow first.
# Replace YOUR_WORKFLOW_ID with the ID in its /flow/ URL.
curl 'https://www.wireflow.ai/api/v1/workflows/YOUR_WORKFLOW_ID/run' \
  -H 'Authorization: Bearer YOUR_WIREFLOW_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{"nodeIds":["model"]}'

# Response (202):
# {"data":{"executionId":"...","status":"running",
#  "poll":"/api/v1/workflows/executions/.../poll"}}
# GET the returned poll path with the same Authorization header.

Open the SDK examples. Read the run API docs or connect your agent.

Read the Seedance 2.5 Flash docs

Working with this model

Seedance 2.5 Flash

Seedance 2.5 Flash is a setting on the Seedance 2.5 card, not a separate node. Set Model to Seedance 2.5 Flash in the sidebar. Over the API or MCP, set quality: "flash" in the Seedance 2.5 node's config.

Flash costs about 35 percent less per second than Seedance 2.5. Credits with no reference video, by length:

Length480p, 2.5480p, Flash720p, 2.5720p, Flash
5 s755013590
10 s150100270180

A run with a reference video holds credits for the longest clip it could make until it finishes, then settles on what the run used, so the difference comes back. With Duration on Auto that is 30 seconds. Duration 5 is the default and holds like Auto, 30 seconds. Any other length holds that length or the reference clip's, whichever is longer. That hold can be larger than Seedance 2.5's for the same short clip: at 720p with a 5 second reference clip and Duration on Auto, Flash holds 860 credits where Seedance 2.5 holds 611, and you need that balance to start the run.

What Flash does: everything the card does (a prompt, a Start Frame and End Frame, reference images, video and audio, native audio, 4 to 30 seconds or Auto), at 480p or 720p.

What it does not do: 1080p, 4k, draft mode, or more than 5 reference videos. Flash takes at most 10 reference images, 5 videos and 5 audio tracks; the card already takes at most 8 images and videos together and 2 audio tracks, so the video cap is the one you can reach. Each of those is refused before you are charged; set Model back to Seedance 2.5 for those shots.

Seedance 2.5 Flash FAQ

What is Seedance 2.5 Flash?

Seedance 2.5 Flash is an AI video generator from ByteDance. The cheaper tier of the Seedance 2.5 card (set Model to Seedance 2.5 Flash).

Is Seedance 2.5 Flash free?

Running it uses Wireflow credits, about 90 credits per 5s, 720p clip. You can set up the inputs here and open a workflow before you spend anything.

How much does Seedance 2.5 Flash cost?

Seedance 2.5 Flash costs about 90 credits per 5s, 720p clip at default settings, and up to about 540 credits for 30 seconds. You pay with Wireflow credits, and the board shows the estimate for your settings before you run.

How long can Seedance 2.5 Flash videos be?

Seedance 2.5 Flash makes clips of 4 to 30 seconds.

What resolution does Seedance 2.5 Flash support?

480p or 720p.

Does Seedance 2.5 Flash make sound?

Yes. It can make audio with the video, and that is on by default.

Can I use Seedance 2.5 Flash commercially?

Yes. What you make with Seedance 2.5 Flash on Wireflow is yours to use for personal and commercial work, subject to the model maker's own terms. Our terms of service have the details.

Does Seedance 2.5 Flash have an API?

Yes. Save a workflow with Seedance 2.5 Flash, then call the Wireflow run API with its workflow ID. The API tab above has examples you can copy, and agents can run it through MCP.

How do I use Seedance 2.5 Flash?

Set the inputs above and choose Open in a workflow. Sign in to save your board, then check its credit estimate before you run.

More AI video models