Multi reference image fusion for video generation

Multi reference image fusion on RelayDance lets you attach up to 9 reference images, 3 reference videos, and 3 audio tracks per request, cited in the prompt as @image1 to @imageN. You submit a task via POST /v1/video/generations, place media in metadata.content[], and poll until the status is succeeded or failed. The API is OpenAI-compatible with base URL https://relaydance.com/v1, and clips run up to 15 seconds.

How reference image fusion works

Reference media are supplied in metadata.content[] and cited in the prompt as @image1 to @imageN. Each request accepts up to 9 reference images, 3 reference videos, and 3 audio tracks, with generated clips up to 15 seconds. To generate a video, send POST /v1/video/generations with model, prompt, seconds, and metadata (ratio, resolution, generate_audio, callback_url, and the content[] reference array). You then poll GET /v1/video/generations/{task_id} until status is succeeded or failed, and the response returns the video url. For asynchronous delivery, set metadata.callback_url so the final state is POSTed to your server. See relaydance.com/docs for details.

Setup and authentication

Setup reuses your existing OpenAI SDK by pointing it at the RelayDance base URL. According to the official relaydance.com/docs, 「Set base_url to https://relaydance.com/v1 and keep your OpenAI SDK」. Authenticate with an Authorization: Bearer YOUR_API_KEY header, creating keys at https://relaydance.com/console. The value here is direct OpenAI-compatible access to Seedance with no BytePlus enterprise account or KYC required. Follow these numbered steps:

1. Create an API key at https://relaydance.com/console.
2. Set base_url to https://relaydance.com/v1.
3. POST /v1/video/generations with model, prompt, seconds, and metadata.content[].
4. Poll GET /v1/video/generations/{task_id} until status is succeeded or failed.
5. Read the returned video url, or use metadata.callback_url for webhook delivery.

Pricing per model

Pricing is pay-as-you-go per generated video, and failed requests carry no cost. According to the official relaydance.com docs, 「Failed or errored requests are never billed」. The table below lists current per-model rates; live values are at relaydance.com/models.

For image billing with gpt-image-2, image output is free and only input is billed, image-to-image starts from about CNY 0.035, and 4K and 1K output cost the same.

Payments and delivery

Payments accept USDT and Stripe card, with billing measured pay-as-you-go per generated video. Since failed or errored requests are never billed, you are charged only for completed outputs. Video results are delivered in one of two ways: synchronous polling via GET /v1/video/generations/{task_id}, which returns the video url once status reaches succeeded, or webhook mode via metadata.callback_url, where the final state is POSTed to your server. Images use POST /v1/images/generations. Available models include ByteDance Seedance 2.0 and Seedance 1.5 Pro, HappyHorse video models, and Grok Imagine video and image models, plus Grok text models. Per-request limits stay fixed at 9 reference images, 3 reference videos, 3 audio tracks, and clips up to 15 seconds.

FAQ

How many reference images can one request use? Up to 9 reference images, alongside 3 reference videos and 3 audio tracks, all placed in metadata.content[] and cited as @image1 to @imageN.

Am I charged for a failed generation? No. 「Failed or errored requests are never billed」, so pay-as-you-go billing applies only to completed videos.

Do I need a BytePlus account or KYC? No. RelayDance provides direct OpenAI-compatible access to Seedance with no BytePlus enterprise account or KYC required.