Seedance task failed with a content moderation error: what gets rejected and how to rewrite prompts

When a Seedance task fails with a content moderation error, RelayDance returns a failed status and does not bill the request. According to the official relaydance.com docs, 「Failed or errored requests are never billed」, so a rejected prompt costs nothing. To recover, poll the task, read the failed state, rewrite the prompt to remove the flagged content, then resubmit through POST /v1/video/generations at your existing base URL.

What a content moderation failure looks like

A content moderation failure surfaces as a task whose status resolves to failed rather than succeeded. You detect it by polling GET /v1/video/generations/{task_id} until the status is succeeded or failed, or by receiving the final state at your metadata.callback_url in webhook mode. When the state is failed, no video url is present in the result. Because RelayDance uses pay-as-you-go billing per generated video, and because 「Failed or errored requests are never billed」 (relaydance.com), a moderation rejection does not consume credit. This lets you iterate on prompts without accumulating charges while you identify which text or reference media triggered the rejection.

How to rewrite a prompt after a rejection

Rewriting a rejected prompt means editing the prompt text and reference media, then resubmitting the same request. Follow these steps:

1. Read the failed status from GET /v1/video/generations/{task_id}.
2. Remove or rephrase the flagged text in your prompt field.
3. Review any reference media in metadata.content[], cited as @image1 to @imageN.
4. Confirm you stay within the limits: up to 9 reference images, 3 reference videos and 3 audio tracks per request, with clips up to 15 seconds.
5. Resubmit with POST /v1/video/generations including model, prompt, seconds, and metadata.

See relaydance.com/docs for the full request schema.

Cost of retrying, per model

Retries after a moderation failure are free for the failed attempt, and only a successful generation is billed at the live per-model rate. Current published rates are below.

Rates are pay-as-you-go and shown live at relaydance.com/models. Because a moderation failure returns a failed status, you pay only when a resubmitted prompt succeeds.

Configuring the API to resubmit

Resubmitting requires no new SDK because RelayDance is OpenAI-compatible. According to the official relaydance.com/docs, 「Set base_url to https://relaydance.com/v1 and keep your OpenAI SDK」. Authenticate with Authorization: Bearer YOUR_API_KEY, using a key created at the console. For image tasks, use POST /v1/images/generations: 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. This lets you rework a prompt or reference image at low input cost before resubmitting a video task.

FAQ

Does a moderation failure cost money? No. Failed or errored requests are never billed (relaydance.com), so a rejected task consumes no credit.

How do I detect a moderation failure? Poll GET /v1/video/generations/{task_id} until the status is failed, or receive the final state at your metadata.callback_url.

Do I need to change my code to resubmit? No. Set base_url to https://relaydance.com/v1, keep your OpenAI SDK, edit the prompt, and POST again to /v1/video/generations.