Troubleshooting

“API Key Required” warning

The app shows this when no API key has been saved yet.

Click the key icon in the top-right header
Select your provider and paste your API key
Click Test to verify the connection

If you’ve already entered a key and still see this, try re-entering it. The key is stored in Windows secure storage — if that’s unavailable on your machine, the app will prompt you each session.

Generation fails with an error

Open the Logs panel (clipboard icon in the header) to see the exact error message per image.

Common causes:

Error	Fix
`401 Unauthorized`	API key is invalid or expired. Re-enter it in settings.
`403 Forbidden`	Your account doesn’t have access to this model or endpoint.
`429 Too Many Requests`	Rate limit hit. The app retries automatically with backoff, but very large batches may exceed your account’s tier. Slow down or upgrade your plan.
`402 Payment Required`	Your account has run out of credits. Add billing at your provider’s dashboard.
`500 / 503 Server Error`	Provider-side error. The app retries automatically. If it persists, try again later.
`content_policy_violation`	The AI rejected the prompt. Review the description for the failed image and rephrase. Avoid references to real people, violence, or explicit content.
`invalid_size`	The template dimensions don’t map to a supported size for this model. For OpenAI: use 1024×1024, 1024×1792, or 1792×1024. For Fireworks FLUX: only 1024×1024 is supported.

Fireworks AI — image looks wrong or wrong size

FLUX models on Fireworks only support 1024×1024 (square). If you’re using a portrait or landscape template, switch to an SDXL model on Fireworks, which supports 768×1344 portrait and 1344×768 landscape.

Some images generated, others failed

Partial failures are normal — AI providers occasionally reject prompts for content policy reasons or experience transient errors. Successfully generated images are saved; failed ones are logged with the reason.

Check the Logs panel for details on which items failed and why. You can fix the offending rows and re-run — skip existing files is on by default so previously generated images won’t be regenerated.

The Generate button is disabled

The button stays disabled until:

An image list file is selected
An output folder is set

Check both fields in Step 3.

Images look nothing like what I described

A few things to try:

Improve the description — be specific about subject, lighting, mood, and composition. Vague descriptions give vague results.
Use the subject and mood fields — these are injected separately and give the model stronger signals.
Check your persona — the visual style in the active persona shapes every image. If the persona has strong stylistic overrides it may override your description. Edit or switch the persona.
Use Prompt Preview — click Preview Prompt in Step 3 to see exactly what’s being sent to the API before spending money.
Switch to HD quality (OpenAI) — for complex scenes, HD often follows the prompt more faithfully.
Try a different content style — if Photorealistic isn’t working, try Cinematic or Digital Art.

Captions not appearing on images

Check that Include Captions is toggled on in Output Settings
Confirm the relevant rows in your image list have a non-empty caption (or text / tagline) column
Use Prompt Preview to verify the caption field is being picked up — the preview shows which fields were parsed

The app won’t start / crashes on launch

Ensure you’re running Windows 10 or later
Try reinstalling from the Microsoft Store
Check for a pending Windows update

If the issue persists, open an issue with a description of what happens.

Output folder not opening

The Open output folder button after generation uses Windows Explorer. If it doesn’t open, navigate manually to your configured output folder and look for the most recent timestamped subfolder (e.g. 2026-03-01_14-30-00).