Skip to main content
This page covers the most common issues encountered during image generation in ZeroTwo Studio and how to resolve them.
Your prompt likely contains content that conflicts with ZeroTwo’s image content policy.How to fix:
  1. Read the error message — it specifies the category of content that was flagged (e.g., explicit content, real person misuse, deceptive representation)
  2. Identify the phrase or element in your prompt that triggered the block
  3. Rephrase or remove that element
  4. Try again with the revised prompt
Most policy blocks are triggered by ambiguous phrasing rather than genuinely prohibited intent. Small changes — like rephrasing “realistic ID document” to “illustrated fictional document” — are often enough to resolve the issue.See Image Safety and Policy for a full list of prohibited content types.
The output quality is closely tied to the specificity and clarity of your prompt and the model you selected.How to fix:
  • Add more detail to your prompt: Include subject description, style, lighting, composition, and mood. A prompt like “a dog” will produce generic results; “a golden retriever puppy in a sunlit meadow, close-up, soft bokeh background, warm natural light” produces a much better result.
  • Specify the style explicitly: Add phrases like “photorealistic photograph”, “watercolor illustration”, or “Studio Ghibli anime style”.
  • Try a different model: Some models follow instructions more precisely than others. GPT-Image-1.5 and Imagen 4.0 are generally the strongest for instruction-following and photorealism. Flux Pro v1.1 is strong for artistic styles.
  • Use a style preset: Select a style from the style dropdown to reinforce the visual direction of your prompt.
See Prompt tips for better images for a detailed guide.
Some models in ZeroTwo require a Pro+ plan and are not available on Free.How to fix:
  • Check your current plan in Settings → Account
  • Premium image generation models (GPT-Image-1.5, Imagen 4.0, Flux Pro, etc.) require Pro or higher
  • Free plan users have access to a limited set of models
  • Upgrade your plan to unlock the full model library
If you are on a Pro+ plan and a model is still grayed out, it may be temporarily unavailable due to provider maintenance. Try again later or select a different model.
Free plan users are limited to 1 image generation per day. This includes all images generated across all models.What to do:
  • Your daily limit resets at midnight UTC
  • Wait until the reset and try again
  • Upgrade to Pro+ in Settings → Account for significantly more monthly generations
  • Note: generating 4 images at once counts as 4 generations against your limit
If you need more than 1 image per day regularly, a Pro plan is the practical solution. The limit on Free is intentional to manage compute costs.
Most image generations complete within 5–30 seconds. High-quality or complex requests can take longer.Normal wait times by scenario:
  • Standard quality, fast model (GPT-Image-1-mini): 5–15 seconds
  • Standard quality, full model: 15–30 seconds
  • HD quality or complex scene: 30–60 seconds
If it exceeds 2 minutes:
  1. Do not close the tab — the generation may still complete
  2. If there is no result after 2–3 minutes, refresh the page
  3. Check if the generation appears in your gallery (it may have completed but the UI didn’t update)
  4. Try generating again — transient slowdowns are usually resolved on retry
ZeroTwo’s image infrastructure may experience higher load during peak usage periods, which can extend generation times.
AI image models can struggle with human faces, especially in complex scenes with multiple subjects, unusual angles, or small face sizes in the frame.How to improve face quality:
  • Use a model optimized for faces: GPT-Image-1, GPT-Image-1.5, and Imagen 4.0 generally render faces better than other models
  • Specify “professional portrait” in your prompt to signal a face-forward composition: "professional portrait, close-up, sharp focus on face"
  • Generate at portrait aspect ratio (9:16): This gives the model more vertical space and often results in the face being larger in the frame
  • Avoid crowds or multiple faces: More than 2–3 people in one image increases the chance of distortion
  • Generate multiple variations: Face rendering varies across generations — generate 4 images and select the best result
Accurate text rendering is a known challenge for AI image models. Most models struggle to generate legible, correctly spelled text within images.How to improve text in images:
  • Use Imagen 4.0: Google’s Imagen model has notably better text rendering than most alternatives — it’s the recommended model when on-image text accuracy is important
  • Keep text short and simple: Single words, short phrases, or simple titles render more accurately than sentences
  • Specify the text clearly in quotes: "a sign that reads 'OPEN'"
  • Generate multiple variations: Text rendering quality varies between generations — try 4 at once and pick the best
  • Accept imperfection for decorative text: If the text doesn’t need to be readable (background signage, decorative elements), the current quality level may be sufficient
For designs where precise text is critical (logos, marketing assets, print), it’s often more reliable to generate the image without text in ZeroTwo and add text separately in a design tool like Figma or Canva.

​
Still having issues?

If none of the above resolved your problem, try these general steps:
  1. Refresh the ZeroTwo page and try again
  2. Clear your browser cache and reload
  3. Try a different browser
  4. Check zerotwo.ai for any active service status notifications
For persistent issues, contact ZeroTwo support with a description of the problem and the prompt you used.