Dictionary issues
Dictionary file not found
Dictionary file not found
ERROR: Dictionary file 'jmdict_enhanced.pkl' not foundCause: The prebuilt dictionary is missing from your installation.Solution:Download prebuilt dictionary
Or build your own
jmdict_enhanced.pkl in your repository root.Dictionary loading is slow
Dictionary loading is slow
- Using JSON dictionary files instead of the pickled format
- Dictionary file is corrupted
- Disk I/O bottleneck (slow hard drive)
- Ensure you’re using
jmdict_enhanced.pkl(pickled format), not JSON files - Rebuild the dictionary if it may be corrupted:
- Move meikipop to an SSD if currently on a slow hard drive
- Check console output for the actual load time: “Dictionary loaded in X.XX seconds”
No results or wrong results
No results or wrong results
- Dictionary version mismatch
- Corrupted dictionary file
- OCR not recognizing text correctly
- First, verify OCR is working by checking what text is being recognized (watch console output with debug logging)
- Try looking up a known common word like 日本 or これ
- Rebuild the dictionary from scratch:
- Check that your meikipop version matches the dictionary format
OCR issues
Google Lens OCR not working
Google Lens OCR not working
- No internet connection
- Firewall blocking requests to Google
- API key issues
- Network timeout
Check firewall settings
Try a different OCR provider
- Right-click system tray icon → OCR Provider → meikiocr (local)
owocr websocket connection failed
owocr websocket connection failed
Could not connect to owocr at ws://127.0.0.1:7331Cause: owocr is not running or not configured correctly.Solution:-
Ensure owocr is installed and up to date:
-
Start owocr with the correct parameters:
-
Verify owocr is running:
- Check that you’re using a compatible owocr backend (glens, oneocr, apple, or bing)
Chrome Screen AI missing components
Chrome Screen AI missing components
Please download the screenai component...Cause: Screen AI library files are not installed.Solution:Download Screen AI components
Extract to correct location
- Linux:
~/.config/screen_ai/resources/ - Windows:
%USERPROFILE%\.config\screen_ai\resources\
Verify files are present
chrome_screen_ai.dll (Windows) or libchromescreenai.so (Linux) is in the resources directory.meikiocr initialization failed
meikiocr initialization failed
failed to initialize meikiocr (local)Possible causes:- meikiocr library not installed
- Model download failed
- Incompatible system configuration
-
Install or update meikiocr:
- Ensure you have sufficient disk space for model downloads (~200MB)
-
Check that your system has:
- CUDA support (for GPU acceleration), or
- CoreML support (macOS), or
- CPU fallback capability
- Check console output for specific error details
- Try running from source to see detailed error messages
Platform-specific issues
macOS: Tray icon not showing
macOS: Tray icon not showing
-
Configure meikipop through
config.inidirectly instead of the GUI: - Use keyboard shortcuts and hotkeys to interact with meikipop
- Check console for any tray-related error messages
macOS: Permission denied errors
macOS: Permission denied errors
Grant Input Monitoring permission
Linux: X11 vs Wayland issues
Linux: X11 vs Wayland issues
-
Check if you’re running Wayland:
-
If it shows “wayland”, switch to X11:
- Log out
- At the login screen, select your username
- Click the gear icon and choose “GNOME on Xorg” (or similar X11 option)
- Log in
- Alternatively, run Xwayland apps in compatibility mode
Windows: Antivirus false positive
Windows: Antivirus false positive
- Add meikipop to your antivirus whitelist/exclusions
- Download from official GitHub releases only
- Verify file integrity if concerned
- Run from source instead of using the packaged executable:
Performance issues
Slow lookup or lag
Slow lookup or lag
- Slow OCR provider
- Network latency (Google Lens)
- Large scan region
- System resource constraints
-
Switch to a faster OCR provider:
- Try meikiocr (local) for fastest local processing
- Enable Google Lens low bandwidth mode if using remote OCR
-
Reduce scan region size:
- Use region mode instead of fullscreen
- Select a smaller area covering just the text you need
- Right-click tray icon → Reselect Region
-
Enable auto scan mode:
- This pre-scans continuously for instant lookups
- Right-click tray icon → Auto Scan Mode
- Close other resource-intensive applications
-
Check console output for timing information:
- Look for “OCR response received in X.XXs”
- Identify which step is slow (OCR vs lookup vs rendering)
High CPU or memory usage
High CPU or memory usage
- Auto scan mode enabled with high frequency
- Large scan region with continuous scanning
- Memory leak (rare)
-
Disable or adjust auto scan mode:
- Use region mode instead of fullscreen scanning
- Reduce scan region to the minimum needed area
- Try a different OCR provider (some are more resource-intensive)
- Restart meikipop if memory usage grows over time
Configuration issues
Settings not saving
Settings not saving
- File permission issues
- config.ini in wrong location
- Syntax errors in config.ini
-
Check that
config.iniexists in the same directory as meikipop - Verify file permissions (should be readable and writable)
-
Look for syntax errors in config.ini:
-
Reset to defaults by deleting
config.iniand restarting meikipop - Check console output for “Settings saved to config.ini” confirmation
Hotkey not working
Hotkey not working
- Hotkey conflict with another application
- Invalid hotkey configuration
- Permissions issue (macOS)
-
Try a different hotkey:
- Check for conflicts with other applications using the same hotkey
- On macOS, ensure Input Monitoring permission is granted
- Restart meikipop after changing hotkey configuration
- Try using auto scan mode as an alternative to hotkey-triggered scanning
Getting help
If your issue isn’t covered here:- Check the console output for error messages and stack traces
- Search existing GitHub issues at github.com/rtr46/meikipop/issues
- Open a new issue with:
- Your operating system and version
- meikipop version
- OCR provider being used
- Relevant console output
- Steps to reproduce the issue