How to generate video with audio through an API

To generate video with audio through an API, send a POST request to /v1/video/generations on RelayDance with a model, a prompt, a seconds value, and metadata that sets generate_audio to true. The service is OpenAI compatible, so you point your existing SDK at the base URL https://relaydance.com/v1. You then poll the task until it returns succeeded or failed, and the result contains the video url. Failed requests are never billed.

Which endpoint and parameters create audio video

Use POST /v1/video/generations with the audio flag enabled in metadata. The request body includes a model, a prompt, a seconds value (clips up to 15 seconds), and a metadata object that can set ratio, resolution, generate_audio, and callback_url. Reference media go in metadata.content[] and are cited in the prompt as @image1 to @imageN. Each request supports up to 9 reference images, 3 reference videos, and 3 audio tracks. Authentication uses an HTTP header: Authorization: Bearer YOUR_API_KEY, with keys created at the RelayDance console. Full parameter details are at relaydance.com/docs.

How to connect with an OpenAI-compatible SDK

You connect by changing only the base URL and keeping your existing OpenAI SDK. According to the documentation, "set base_url to https://relaydance.com/v1 and keep your OpenAI SDK"(According to the official relaydance.com/docs). This means no separate BytePlus enterprise account or KYC is required to reach Seedance models. The steps are: 1) create an API key at the console, 2) set base_url to https://relaydance.com/v1, 3) submit a video task with your prompt, seconds, and metadata, 4) poll GET /v1/video/generations/{task_id} until status is succeeded or failed, and 5) read the video url from the result.

How polling and webhook delivery work

You retrieve results either by polling the task endpoint or by registering a webhook callback. To poll, call GET /v1/video/generations/{task_id} repeatedly until the status is succeeded or failed; the succeeded result contains the video url. For webhook mode, set metadata.callback_url, and the final state is POSTed to your server, removing the need to poll. Both modes use the same task submission flow described above. Payment is pay-as-you-go per generated video, with USDT and Stripe card accepted. Regarding errors, "failed or errored requests are never billed"(According to the official relaydance.com docs), so polling a failed task does not add cost.

How much does video generation cost

Pricing is per second or per fixed bar, depending on the model and resolution. The numbers below are drawn from relaydance.com/models, where live per-model rates are published.

For images, gpt-image-2 charges only for input: image output is free, image-to-image starts at about CNY 0.035, and 4K output is priced the same as 1K output.

FAQ

• Am I charged for a failed request? No. Billing is pay-as-you-go per generated video, and "failed or errored requests are never billed"(According to the official relaydance.com docs).
• Do I need a BytePlus enterprise account or KYC? No. RelayDance provides OpenAI-compatible access to Seedance with no BytePlus enterprise account or KYC required.
• How many reference files can one request use? Up to 9 reference images, 3 reference videos, and 3 audio tracks, with clips up to 15 seconds.