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, the request returns a failed status and is not billed, so you can safely adjust the prompt and resubmit. According to relaydance.com/docs, "failed or errored requests are never billed"(relaydance.com). Poll the task with GET /v1/video/generations/{task_id} until status reads failed, then rewrite the prompt or reference media (@image1 to @imageN) and POST a fresh task to /v1/video/generations.

What billing applies when a task is rejected

A moderation rejection carries no cost because failed or errored requests are never billed. RelayDance uses pay-as-you-go pricing per generated video, so a task that ends in the failed state does not consume credit. Per relaydance.com official docs: "failed or errored requests are never billed"(relaydance.com). For reference, successful generations are priced separately: Seedance 2.0 720p is about $0.190 per second and 1080p is about $0.470 per second (source relaydance.com/models). This means you can iterate on a rejected prompt and only pay once a task reaches the succeeded state.

How to detect the failure and resubmit

Detect a moderation failure by reading the task status, then resubmit a corrected request through the same OpenAI-compatible interface. According to relaydance.com/docs, "set base_url to https://relaydance.com/v1 and keep your OpenAI SDK"(relaydance.com/docs). Follow these steps:

1. POST your task to /v1/video/generations with model, prompt, and seconds.
2. Poll GET /v1/video/generations/{task_id} until status is succeeded or failed.
3. If status is failed, review the prompt text and each metadata.content[] reference.
4. Rewrite the prompt, then POST a new task; the rejected attempt is not billed.

How to rewrite prompts and reference media

Rewrite prompts by removing the flagged element and re-checking each reference cited as @image1 to @imageN. Requests support up to 9 reference images, 3 reference videos, and 3 audio tracks, with clips up to 15 seconds, so a single flagged reference can be swapped without rebuilding the whole request. Reference media go in metadata.content[] and are cited in the prompt by their @image index. After editing, resubmit through base_url https://relaydance.com/v1 with your existing OpenAI SDK. Since Seedance Fast is about $0.152 per second and native 4K is about $4.90 per 5 second bar (source relaydance.com/models), choosing a lower tier during iteration limits cost once a corrected task succeeds.

Cost comparison for iteration tiers

Select a tier that matches your iteration budget, since rejected attempts cost nothing and only succeeded tasks are billed.

FAQ

Am I charged when Seedance rejects my prompt? No. Failed or errored requests are never billed, so a moderation rejection carries no cost.

How do I confirm a task failed? Poll GET /v1/video/generations/{task_id} until the status reads failed, or use webhook mode by setting metadata.callback_url so the final state is POSTed to your server.

Do I need a new integration to resubmit? No. Change base_url to https://relaydance.com/v1 and keep your OpenAI SDK, then POST the corrected task to /v1/video/generations.