How cuemeet/cuemeet-google-bot works
Project type and purpose
• This is a full-stack automation/integration tool: a Google Meet “bot” service combined with a browser extension for transcript capture.
• Purpose: programmatically join Google Meet calls, stay “alive” (random mouse movement), capture transcripts via a Chrome extension, and expose control via a backend API.
File structure
• Top-level service:
– app.py: main Flask/FastAPI (or similar) web service entry point
– config/: settings module for environment, secrets, and feature toggles
– logger.py & monitoring.py: centralized logging and metrics instrumentation
– google_meet/: core bot logic (bot.py orchestrates browser automation; random_mouse.py for anti-idle; utils.py for common helpers)
– transcript_extension/: Chrome extension assets (manifest.json, background.js, content.js)
– Dockerfile & entrypoint.sh: containerization
– buildspec*.yml: CI/CD build definitions
– pulse-*.conf: audio subsystem (PulseAudio) configurations
• Supporting scripts: Makefile, .env.example, pyproject.toml
README for additional insights
• Confirms this is one of several meeting-bots (Zoom, Teams, etc.), focused on Google Meet
a. Main components
1. User’s Browser + Google Meet UI
2. Transcript Chrome Extension (content.js + background.js)
3. Backend Service (app.py + google_meet package)
4. Browser Automation Engine (e.g., Selenium or Puppeteer, invoked by bot.py)
5. Audio Subsystem (PulseAudio with pulse-client.conf)
6. Logging & Monitoring (logger.py, monitoring.py + external metrics/log aggregator)
7. External Service: Google Meet Servers
b. Relationships & interactions
1. The user navigates to Google Meet in their browser; the extension injects content scripts.
2. content.js scrapes transcript text/events, sends them to background.js.
3. background.js POSTs transcript data to the backend’s REST/WebSocket endpoints.
4. The backend (app.py) exposes control APIs: start/stop meeting bot sessions, configure settings.
5. When a session is started, bot.py spins up a headless browser, connects to Google Meet, runs random_mouse to simulate activity, uses PulseAudio for voice or audio capture if needed.
6. Bot interactions (join/leave, transcript forwarding) are logged via logger.py and metrics emitted via monitoring.py.
7. Logs/metrics are sent to an external logging/monitoring service (e.g., CloudWatch, Prometheus+Grafana).
8. The CI/CD pipeline (buildspec.yml) builds the Docker image via Dockerfile and deploys to staging/production.
c. Architectural patterns & design principles
• Separation of concerns:
Frontend (extension) vs. Backend (API + automation) vs. Infrastructure (logging/monitoring)
• Modularization: google_meet module encapsulates bot logic; transcript_extension is fully decoupled.
• Containerization: Docker + buildspec for reproducible builds and deployments
• Configuration-as-code: config/settings.py + .env for environment-specific overrides
d. Key technologies/libraries
• Python 3.x (pyproject.toml)
• Flask or FastAPI (inferred from app.py)
• Selenium or Puppeteer (browser automation in bot.py)
• Chrome Extension APIs (manifest.json, content/background scripts)
• PulseAudio for audio routing
• Docker, AWS CodeBuild (buildspec), environment variables