Troubleshooting
Common issues and solutions for GUI automation.
VM Issues
VM Won’t Start
# Check Docker
docker ps -a | grep osworld
# View logs
docker logs osworld-vm
# Restart container
docker restart osworld-vm
Permission Issues (macOS)
Accessibility Permission
- Open System Settings
- Privacy & Security → Accessibility
- Add Terminal/Python
- Restart application
Screen Recording Permission
- Open System Settings
- Privacy & Security → Screen Recording
- Add Terminal/Python
- Restart application
Model Issues
API Call Failed
Check:
- ✅ API Key correct
- ✅ Network connection OK
- ✅ API quota sufficient
Qwen-VL Test:
curl -X POST https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions \
-H "Authorization: Bearer $DASHSCOPE_API_KEY" \
-d '{"model": "qwen3-vl-plus", "messages": [{"role": "user", "content": "test"}]}'
Performance Issues
Slow Response
Solutions:
- Use faster model (qwen3-vl-flash)
- Reduce max_steps
- Lower screenshot resolution
Actions Not Executing
Check:
- Verify coordinates are within screen bounds
- Ensure target window is visible
- Check accessibility permissions