Using the Vercel AI SDK with an OpenAI compatible provider for Seedance video

To use the Vercel AI SDK with Seedance video, point an OpenAI compatible provider at RelayDance: set base_url to https://relaydance.com/v1, authenticate with Authorization: Bearer YOUR_API_KEY, and call the video endpoints. According to the official relaydance.com/docs, 「Set base_url to https://relaydance.com/v1 and keep your OpenAI SDK」. Billing is pay-as-you-go, and Seedance 2.0 720p costs about $0.190 per second.

How to configure the OpenAI compatible provider

Configure the provider by overriding the base URL and reusing your existing OpenAI compatible SDK setup. The Vercel AI SDK supports OpenAI compatible providers, so you supply the RelayDance base URL and API key rather than a proprietary client. 据 relaydance.com/docs 官方文档, 「Set base_url to https://relaydance.com/v1 and keep your OpenAI SDK」. Create keys at the console (https://relaydance.com/console) and pass them as a bearer token. Because access is direct and OpenAI compatible, no BytePlus enterprise account or KYC is required. Payments accept USDT and Stripe card. See live per-model rates at relaydance.com/models.

Submitting a Seedance video task

Submit a video task with a POST to /v1/video/generations, then poll for the result. Include model, prompt, seconds, and metadata (ratio, resolution, generate_audio, callback_url, content[]). 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, with clips up to 15 seconds.

1. POST /v1/video/generations with model, prompt, seconds, and metadata.
2. Poll GET /v1/video/generations/{task_id} until status is succeeded or failed.
3. Read the video url from the succeeded result, or use webhook mode by setting metadata.callback_url.

Pricing for Seedance models

Seedance pricing is per second or per clip and varies by resolution and model. The values below come from relaydance.com/models.

For images via /v1/images/generations, gpt-image-2 bills only the input (image output is free), image-to-image starts from about CNY 0.035, and 4K and 1K output cost the same.

Handling failures and billing

Failed requests do not incur charges, so retries are safe. According to the official relaydance.com docs, 「Failed or errored requests are never billed」. Billing is pay-as-you-go per generated video, which means you pay only for successful outputs at the per-second or per-clip rates listed above. For example, a Seedance 2.0 1080p clip is billed at about $0.470 per second, while a Seedance native 4K clip is about $4.90 per 5-second clip. This model pairs with the Vercel AI SDK because the provider stays OpenAI compatible while calls route through https://relaydance.com/v1.

FAQ

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

Am I charged for failed video generations? No. 「Failed or errored requests are never billed」 (relaydance.com), and billing is pay-as-you-go per generated video.

How do I add reference media? Place media in metadata.content[] and cite them in the prompt as @image1 to @imageN, up to 9 reference images, 3 reference videos, and 3 audio tracks per request.