Troubleshooting

Database Issues

If the database becomes corrupted or you want a fresh start:

rm ~/.coodeen/data.db
npx coodeen

A new database will be created automatically on the next launch.

AI Can't See Screenshots

Make sure you're using a vision-capable model:

• OpenAI: GPT-4o, GPT-4.1
• Anthropic: Claude Sonnet 4, Claude Opus 4
• Google: Gemini 2.5 Pro, Gemini 2.5 Flash

Models like GPT-4.1-nano don't support image inputs.

Port Already in Use

If port 3099 is already in use, Coodeen will fail to start. Kill the existing process:

lsof -ti:3099 | xargs kill
npx coodeen

Preview Not Loading

If the preview panel shows a blank page:

1. Make sure your dev server is actually running on the URL you entered
2. Check for CORS issues — some frameworks block iframe embedding
3. Try reloading the preview with the refresh button
4. Check the browser console for errors

API Key Not Working

1. Verify the key is correct in Settings
2. Check that the key has the necessary permissions (some providers require specific access levels)
3. Ensure you have billing set up with the provider
4. Try regenerating the key from the provider's dashboard

Slow Responses

Response speed depends on:

• Your chosen AI provider and model
• The complexity of the request
• Network latency to the provider's API
• The provider's current load

Smaller models (GPT-4.1-mini, Claude Haiku) are faster but less capable. Larger models (GPT-4.1, Claude Opus 4) are more capable but slower.