Auto-Meet-Join: Automated Meeting Transcription & Speaker Diarization

Features

• Auto-join Google Meet, Zoom, or Teams meetings
• Record system audio via VB-Cable
• Transcribe with OpenAI Whisper
• Speaker diarization and mapping with pyannote-audio
• Speaker-attributed transcript output
• OAuth integration for Zoom and Teams APIs
• Web-based configuration management

Setup

1. Environment Setup

python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt

if you can't execute this instructions,

Remove-Item -Recurse -Force .\venv
python -c "import compileall; compileall.compile_dir(r'C:\Users\admin\AppData\Local\Programs\Python\Python313\Lib', force=True)"

and try again.

2. API Credentials Setup

Zoom API Setup

1. Go to Zoom App Marketplace
2. Create a new app with "OAuth" type
3. Set redirect URL to: http://localhost:5000/zoom/callback
4. Copy Client ID and Client Secret
5. Use the web interface to configure credentials

Microsoft Teams API Setup

1. Go to Azure Portal
2. Register a new application in Azure AD
3. Add redirect URI: http://localhost:5000/teams/callback
4. Grant permissions: User.Read, OnlineMeetings.ReadWrite
5. Create a client secret
6. Copy Client ID, Client Secret, and Tenant ID
7. Use the web interface to configure credentials

HuggingFace Token

1. Go to HuggingFace Settings
2. Create a new token with read permissions
3. Use the web interface to configure the token

3. Audio Setup

• Download and place nircmd.exe in the project root (for Windows audio device management)
• Place reference speaker samples (e.g., alice.wav, bob.wav) in the root
• Set VB-Cable as the default input device in Windows sound settings

Usage

Web Interface

1. Start the application:

   python app.py

2. Open browser to http://127.0.0.1:5000/
3. Click "Setup API Keys" to configure credentials
4. Enter meeting link and select platform
5. Click "Start Transcription"

Command Line

python auto_transcribe_meeting.py [meet|zoom|teams] <meeting_link>

Configuration

• All API credentials are stored in config.json
• Use the web interface at /setup to manage credentials
• OAuth tokens are stored in session for secure access

Notes

• Set VB-Cable as the default input device for system audio recording
• Requires HuggingFace token for pyannote-audio diarization
• Output transcript is saved to transcript.txt
• Full meeting audio is saved to meeting.wav for diarization
• OAuth tokens expire and need periodic refresh