Black Friday: Up to 66% off most AI models.

Explore the deals

Prompt and settings

Prompts, references and generation modes.

Settings or inputs not accepted

A parameter value or input doesn't match what this model accepts, for example an unsupported aspect ratio, a missing required image, or a mode that needs a reference file.

How to fix it

  1. Click Edit on the failed card to restore your inputs, then check each option.
  2. Make sure the images or videos required by the selected mode are uploaded.
  3. Set the video length with the duration setting, not with commands in the prompt.
  4. If the model rejects an uploaded image, try a different one.
Original messages you may see
  • Prompt is invalid or rejected by provider, please try a different input.
  • Some required request parameters are missing. For more information, see the API document.
  • The prompt contains an invalid duration instruction.
  • Your input was rejected, please try again with different inputs.

Prompt too long

The prompt is longer than this model allows. Models count length differently, in characters or tokens, so one model's limit doesn't apply to another.

How to fix it

  1. Keep the subject, action, style and key visual details, and remove repeated descriptions.
  2. Shorten it to fit the current model, then add details back gradually if needed.
Original messages you may see
  • Prompt exceeds provider limit (max 2000 chars)
  • The prompt is too long for this model. Please shorten it and try again.
  • PromptLengthExceeded

Duplicate images or missing image references

The upload list contains the same image twice, or the prompt refers to an image number (such as @Image 2) that was not uploaded.

How to fix it

  1. Remove duplicate reference images.
  2. Check that every image number in the prompt matches the current uploads.
  3. Upload the missing files, or delete references to images that are no longer there.
Original messages you may see
  • There are duplicate images, please try different pictures.
  • Invalid image reference: @image2, available images: 1-1, please try different prompt.

Settings don't fit the generation mode

In video editing or extension, the output ratio and duration follow the input video, so the usual generation settings may not apply. Some extra features, such as web search, only work in specific modes.

How to fix it

  1. Check whether you are generating, editing or extending a video.
  2. Use the automatic ratio or duration when the mode offers it.
  3. Turn off extra features that the current mode does not support.
Original messages you may see
  • Seedance identified your task as video editing based on your prompt.
  • web_search button only can be used in the mode of text-to-video, please try again different inputs.

Lyrics or preview text not accepted

Music extension and voice preview have their own input rules for lyrics and preview text.

How to fix it

  1. For music extension, check that the lyrics are not empty, too short or badly formatted.
  2. For voice preview, provide preview text that is long enough.
Original messages you may see
  • Task failed. Possible causes: the extending lyrics are empty, too short, or malformed. Please check the lyrics field and retry.
  • Preview_text should not be shorter than 15 characters,please try different again.