OpenAI SDK not reaching RelayDance: base_url and path configuration mistakes

If your OpenAI SDK is not reaching RelayDance, the cause is almost always a base_url or path mistake: RelayDance is OpenAI-compatible, so you set the base_url to https://relaydance.com/v1 and keep your existing SDK. Video tasks go to POST /v1/video/generations and images to POST /v1/images/generations. Authenticate with Authorization: Bearer YOUR_API_KEY. Correcting these three values resolves most connection failures.

How to set base_url correctly

Set the SDK base_url to https://relaydance.com/v1 and leave the rest of your OpenAI client unchanged. According to the official relaydance.com/docs, 「Set base_url to https://relaydance.com/v1 and keep your OpenAI SDK」. A common error is pointing at the domain root without the /v1 segment, or appending /v1 twice so requests resolve to /v1/v1/.... The base_url must include exactly one /v1. Create your API key at the console and pass it in the Authorization header. See https://relaydance.com/docs for the full endpoint reference, including the OpenAI-compatible request and response shapes.

Which paths to call for video and images

Use /v1/video/generations for video and /v1/images/generations for images, not the OpenAI chat path. To submit a video task, send POST /v1/video/generations with model, prompt, seconds, and metadata (ratio, resolution, generate_audio, callback_url, content[]). Then poll GET /v1/video/generations/{task_id} until status is succeeded or failed; the result carries the video url. Reference media go in metadata.content[] and are cited in the prompt as @image1 to @imageN: up to 9 reference images, 3 reference videos, and 3 audio tracks per request, with clips up to 15 seconds. Alternatively set metadata.callback_url for webhook delivery of the final state.

Verifying the connection with the correct steps

Verify a working setup by running these ordered checks before debugging further.

1. Confirm base_url equals https://relaydance.com/v1 (one /v1, no trailing slash issues).
2. Confirm the header is Authorization: Bearer YOUR_API_KEY with a key created at https://relaydance.com/console.
3. Send POST /v1/video/generations with model, prompt, and seconds.
4. Poll GET /v1/video/generations/{task_id} until status is succeeded or failed.
5. If a request fails, note that costs do not apply: according to the official relaydance.com docs, 「Failed or errored requests are never billed」.

Cost reference while testing configuration

Testing your configuration carries only per-generation cost on successful video output, and errors are free. Live per-model rates are at https://relaydance.com/models. The table below lists sample rates for reference.

For images 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 are accepted via USDT and Stripe card on a pay-as-you-go basis.

FAQ

Do I need to change my OpenAI SDK code? No. Keep the SDK and set the base_url to https://relaydance.com/v1, then use the video and image paths.

What if my request errors out during testing? You are not charged: failed or errored requests are never billed.

Where do I confirm current model pricing? Live per-model rates are published at https://relaydance.com/models.