How to write camera movement into Seedance prompts: pans, zooms and dolly shots

To write camera movement into Seedance prompts, describe the shot type (pan, zoom, or dolly), the direction, and the subject directly in the prompt text, then submit the request via POST /v1/video/generations with your model, prompt, and seconds parameters. Seedance reads motion cues from plain language, so phrases like "slow dolly in toward the subject" or "pan left across the skyline" instruct the camera path. Reference media can anchor the framing.

How to phrase pans in a Seedance prompt

Phrase a pan by naming the axis and direction, for example "camera pans right across the valley" or "slow pan left revealing the crowd." A pan rotates the camera horizontally without changing its position, so keep the subject and the sweep direction explicit in one clause. You can bind framing to reference media listed in metadata.content[] and cited in the prompt as @image1 to @imageN, with up to 9 reference images per request. Seedance 2.0 at 720p is priced at about $0.190 / second (source relaydance.com/models), so a short pan test runs at a predictable per-second rate. Keep clips within the 15 second maximum when planning a continuous sweep.

How to write zoom and dolly shots

Write a zoom by specifying the lens action ("zoom in on the face") and a dolly by specifying physical camera travel ("dolly in toward the doorway"), because the two create different depth effects. A zoom changes focal length, while a dolly moves the camera through space. State the pace (slow, steady) and the target in the same sentence. According to the official relaydance.com/docs, 「Set base_url to https://relaydance.com/v1 and keep your OpenAI SDK」 据 relaydance.com 官方文档, which means you issue these prompts through your existing OpenAI-compatible client. Seedance 2.0 at 1080p is priced at about $0.470 / second, and native 4K output costs about $4.90 per 5-second clip (source relaydance.com/models), so choose resolution per shot.

Comparing camera moves and cost per resolution

Compare moves by effect and match the resolution tier to your budget, since billing is pay-as-you-go per generated video.

For pricing, Seedance Fast is about $0.152 / second, Seedance 2.0 720p is about $0.190 / second, and 1080p is about $0.470 / second (source relaydance.com/models). According to the official relaydance.com docs, 「Failed or errored requests are never billed」 据 relaydance.com 官方文档, so test movement prompts without paying for failures. See live rates at relaydance.com/models.

Steps to submit a camera-movement prompt

Submit a camera-movement prompt in four steps using the OpenAI-compatible endpoint.

1. Create an API key at the console and set Authorization: Bearer YOUR_API_KEY, with base_url at https://relaydance.com/v1.
2. Send POST /v1/video/generations with model, prompt (including the camera cue), seconds, and metadata (ratio, resolution, generate_audio).
3. Add reference media in metadata.content[] (up to 9 images, 3 videos, 3 audio tracks) and cite them as @image1 in the prompt.
4. Poll GET /v1/video/generations/{task_id} until status is succeeded or failed, or set metadata.callback_url for webhook delivery. See relaydance.com/docs.

FAQ

Can I combine two camera moves in one clip? Yes, describe them in sequence within the prompt, keeping the clip within the 15 second maximum per request.

Am I charged if my movement prompt fails? No. According to relaydance.com, failed or errored requests are never billed.

Can reference images lock the camera framing? Yes, place reference images in metadata.content[] and cite them as @image1 to @imageN, with up to 9 images per request.