AI-generated images and videos can help creators produce stunning visuals quickly, but users sometimes encounter common issues during generation. This article provides solutions to frequent problems users encounter when generating AI images and videos, including video glitches, gibberish text, and error messages.
I’m trying to create a video with a choreographed dance. The video output is a mess. How can I fix this?
Complex motion often creates visual glitches.
Recommendations:
- Humans: Use small, realistic gestures like looking over a shoulder, shifting weight, or lightly waving a hand.
Animals: Stick to natural actions like walking, sitting, or turning the head.
AI handles smooth, minimal actions more reliably than chaotic or highly detailed movements. This results in cleaner and more realistic video output.
Avoid: Backflips, complicated dance moves, or full-body acrobatics.
I received the error message, “image/video not generated due to inappropriate content. No credits were used.” What does this mean?
This indicates your prompt contains content that violates safety or model requirements (e.g., NSFW content, profanity, or nudity). A seemingly safe image may be picked up as alluding to something NSFW and may be blocked by the filter. Errors are automatic and may not be 100% accurate.
Content the violates guidelines includes:
- Content depicting NSFW content, promoting illegal acts or hate speech.
- Severely harmful content, such as depictions of extreme violence, gore, or content promoting self-harm.
- Content intended to promote misinformation, deception, harassment, or discrimination.
- Generation of content that infringes on third-party intellectual property rights.
- Content that perpetuates harmful stereotypes.
Solution: Edit your prompt or upload a new image and try again. No credits were used for this attempt.
I received the error message, “Something went wrong. No credits were used. Please try again.” What does this mean?
This is a generic error for unknown or temporary failures. This could also relate to the file or image uploaded being too large. Files must be under 10MB and the ideal size for an image is 1 MP or higher (e.g. 1280×720 px).
Solution: Retry your generation later and / or with a different model. If the issue persists, contact support. No credits were used.
I received the error message, “Couldn’t generate image/video. The image you uploaded has expired. No credits were used. Please re-upload and try again.” What does this mean?
This means the uploaded image expired due to inactivity.
Solution: Re-upload the image and try again. No credits were used.
I received the error message, “Couldn’t generate video. Please follow these guidelines when creating with the Sora 2 model: Don’t upload images with real people or faces. Don’t prompt for celebrities or characters. Don’t include explicit, violent, or harmful content.” What does it mean?
The Sora 2 model restricts:
- Images containing real people or faces
- Prompts for celebrities or characters
- Explicit, violent, or harmful content
Solution: Adjust your image or prompt to comply. No credits were used.
The text on my image/video is gibberish. How can I fix this?
The first step is to include quotation marks around the text you want to include in your prompt.
Recommended: “Write “WELCOME HOME” on the billboard in bold white letters.”
Avoid: "Generate a Billboard with the words welcome home."
Keep the amount of text small. The fewer words (especially short phrases rather than paragraphs) the higher the chance of legibility. Avoid punctuation, mixed case, or complex phrases. Simple uppercase works best.
The face and hands in my output look unnatural. How can I fix this?
Unnatural faces or hands often result from incomplete prompt details or low resolution.
Solutions:
- Describe age, pose, lighting, and people in detail.
- Use Image-to-Image or Image-to-Video to provide reference images.
- Increase resolution: Higher resolution reduces distortions.
Why is there an error when uploading a transparent image?
Currently, transparent elements are not supported by AI Image-to-Image or Image-to-Video generators.
Solution: Save the file as a JPEG instead of PNG and re-upload.
I’m getting an error message that says my image is transparent, but it isn’t transparent. What should I do?
Even if the image doesn’t look transparent, it might still contain a hidden transparency layer (called an “alpha channel”), which our system detects as transparency.
Solution:
- Open the file in Preview (Mac) or Paint (Windows).
- Go to File → Export / Save As.
- Save it as a JPEG, or if saving as a PNG, uncheck the Alpha option (if available).
This removes the hidden transparency layer and usually resolves the upload issue.