How screen monitoring works
Screen Answerer uses your browser’s native screen capture API to:- Capture screenshots of your selected window/screen every 5 seconds
- Send the image to Google Gemini API for analysis
- Detect if the image contains a quiz question
- Generate and display the answer if a question is found
The monitoring interval is set to 5 seconds to balance responsiveness with API quota management. This prevents excessive API calls while still detecting questions quickly.
Browser compatibility
Screen monitoring works best with:- Chrome/Chromium - Full support, recommended
- Firefox - Full support with adjusted frame rate
- Edge - Full support
- Safari - Limited support (may have restrictions)
Starting the screen monitor
Ensure API key is configured
Before starting the monitor, make sure you’ve added your Google Gemini API key:
- Click the ⚙️ Settings icon
- Navigate to the API Key tab
- Enter and save your API key
If you haven’t set up an API key yet, see the Getting API Key guide.
Click 'Start Monitoring'
In the Screen Monitor section of the app, click the Start Monitoring button.Your browser will prompt you to select what to share:
- Entire Screen - Monitors your entire display
- Window - Monitors a specific application window
- Browser Tab - Monitors a single browser tab
Grant screen capture permission
Click Share in the browser permission dialog to allow Screen Answerer to capture your screen.
Monitor starts running
Once permission is granted:
- The status changes to Active (displayed in purple)
- The monitor status shows “Monitoring active - watching for quiz questions”
- A preview of the captured screen appears below
- The Stop Monitoring button becomes enabled
When a question is detected
When Screen Answerer identifies a quiz question:- Answer Display - The answer appears in the ANSWER section at the top
- Results Section - The full response is shown in the results area below
- Counter Updates - The “Questions Processed” counter increments
- Status Message - Shows “Quiz question detected! Answer displayed above.”
Screen Answerer displays only the most relevant answer (the first response from Gemini). This keeps the interface clean and focuses on the correct answer.
Stopping the monitor
To stop monitoring:- Click the Stop Monitoring button
- The status changes back to Inactive
- The monitor status shows “Monitoring stopped”
- Screen capture permission is released
If you close the tab or stop sharing from your browser’s screen sharing controls, monitoring will automatically stop.
Understanding the interface
Status indicators
- Status - Shows “Active” (purple) when monitoring, “Inactive” (default color) when stopped
- Questions Processed - Running count of questions detected during the current session
- Monitor Status - Text description of current activity
- Preview - Live thumbnail of what’s being captured
Answer display
Answers are shown in two places:- ANSWER section (top) - Quick reference for the current answer
- Results section (bottom) - Full formatted response with markdown rendering
Rate limiting and performance
Screen Answerer includes built-in rate limiting to protect your API quota:- Client-side throttling - Minimum 5 seconds between requests
- Server-side limiting - Maximum 100 requests per 15 minutes per IP
- API quota tracking - Monitors approach to daily/minute limits
Tips for effective monitoring
Choose the right capture area
Choose the right capture area
Window capture is more efficient than full screen:
- Reduces image size
- Faster processing
- Uses less API quota
- Better focus on relevant content
Position questions clearly
Position questions clearly
For best detection:
- Ensure text is readable and not obscured
- Avoid overlapping windows
- Use adequate screen brightness
- Maximize the quiz window for clearer capture
Monitor API usage
Monitor API usage
Keep an eye on the Questions Processed counter:
- High counts may indicate the detector is too sensitive
- If non-questions are detected, adjust window selection
- Consider stopping monitoring when not actively taking quizzes
Use appropriate model
Use appropriate model
Choose the right Gemini model in settings:
- gemini-2.0-flash-lite (default) - Faster, lower cost, good accuracy
- gemini-2.0-flash - More capable, slightly slower, better for complex questions
Privacy and security
Screen Answerer handles your data securely:- Temporary storage - Screenshots are deleted immediately after processing (see
history.safelyDeleteFile()in server.js:229) - No logging - Images are not stored on Screen Answerer’s servers
- Direct API calls - Images go directly to Google Gemini; Screen Answerer doesn’t intercept them
- Local API key - Your API key stays in your browser’s localStorage
Troubleshooting
Monitor won’t start
- Check API key: Ensure a valid key is saved in Settings
- Browser permissions: Verify screen capture permission is granted
- Browser support: Try Chrome or Firefox if using a different browser
Questions not detected
- Image quality: Ensure text is clear and readable
- Capture area: Verify the question is within the captured region
- Detection interval: Wait at least 5 seconds after question appears
- Model selection: Try switching to
gemini-2.0-flashfor better detection
API errors
- “Invalid API key”: Re-enter your API key in Settings
- “Quota exceeded”: Wait for quota reset (check Google Cloud Console)
- “Rate limited”: Reduce monitoring frequency or wait a few seconds
Performance issues
- High CPU usage: Select a specific window instead of entire screen
- Slow responses: Check your internet connection
- Browser lag: Close unnecessary tabs and applications
Next steps
Best practices
Optimize your screen monitoring workflow
Model selection
Learn which Gemini model to use