Seeing an error when using Descript’s AI tools to generate images or videos? These errors come directly from our model providers. Here’s what they mean and how to fix them.
Image generation rejected based on content
This message tends to occur when generating videos with the AI Video Maker.
In your AI video, you’ll see a checkered placeholder layer with the message “Custom prompt required” or “Image generation rejected based on content. Try again with a custom prompt.”
This happens when Descript sends a prompt to the image generation model and that prompt is rejected by the provider's content filters.
How to replace the placeholder layer:
- Click the placeholder layer in the editor.
- Click the Add Media button in the hover menu.
- Select a media source — upload from your computer, or generate a new image using your own prompt.
- The new media will replace the error message in your video.
The content could not be processed because it contained material flagged by a content checker
This error appears when your image or video prompt (or input image) violates the model’s content safety guidelines. Each provider includes their own content filters to flag inappropriate or policy-violating material.
How to fix it:
- Review your prompt and remove any potentially sensitive language. If using an input image, try swapping in a different reference image.
- Simplify your prompt to use neutral, descriptive language.
- If the error persists with appropriate content, try removing the reference image or rewriting the prompt entirely.
An internal server error occurred
This error commonly appears when using the Sora 2 video model to convert an image containing a human face into a video. Sora 2 can’t process image-to-video generation when the input includes a human face — this is a known limitation of the model.
How to fix it:
- Option 1: Switch to another model like Kling O1, Veo 3.1, or Pixverse v5, which can handle human faces. Learn more about generative image and video models in Descript.
- Option 2: Use a text-only prompt instead of an image input. Sora 2 can generate videos with human subjects from text, just not from images of real faces.
Unable to generate video. This may be due to content restrictions.
This is a general error that can occur for several reasons: rate limiting, temporary service outages, content policy violations, etc.
How to fix it:
- Wait and retry: If you've been generating a lot of video in quick succession, wait a few minutes and try again.
-
Simplify your script: If you're using AI Video Maker:
- Remove section headers, speaker labels, and bracketed notes
- Delete extra spaces or special characters
- Try translating non-English scripts to English
- Modify your prompt: Rephrase it using simpler, more descriptive language.
- Try a different approach: First generate a still image; then use your impact as a reference input to generate a video.