MeetingMind-AI is a cutting-edge AI-powered system designed to automate capturing, summarizing, and communicating business meeting insights with efficiency and professionalism. From raw audio recordings, the system delivers accurate transcriptions, extracts critical meeting highlights, sentiments, and action items, then seamlessly communicates outcomes via email.
- Automated Audio Transcription: Converts meeting audio files into accurate text using advanced speech recognition models like Mistral AI.
- AI-Powered Summarization: Distills lengthy discussions into concise, actionable meeting minutes including key financial highlights and growth strategies.
- Sentiment Analysis: Evaluates tone and sentiment to provide context to meeting outcomes.
- Persistent Knowledge Output: Generates detailed markdown files documenting summaries, action points, and sentiments.
- Direct Gmail Integration: Automatically sends professional emails with meeting minutes to stakeholders.
- Robust Error Handling: Handles API limits and encoding issues gracefully.
- Python 3.8+
- CrewAI Framework
- Mistral AI (Speech-to-Text; Language Models)
- Gmail API
- Pydantic (Data Validation and Schemas)
- Pydub (Audio Processing)
- Python 3.8+ installed
- Mistral AI API key
- Google API credentials for Gmail access
git clone https://github.com/devloper-student/MeetingMind-AI.git
cd MeetingMind-AI
pip install -r requirements.txtCreate a .env file and add the following:
MISTRAL_API_KEY=your_mistral_api_key_here
Place your meeting audio file as EarningsCall.wav in the project root directory.
Run the main program:
python src/main.pyThe system will generate transcripts, create summaries, write markdown meeting minutes files, and send emails to your configured recipients.
Audio Input �1 Transcription �1 AI Analysis �1 Document Generation �1 Email Delivery
�1 �1 �1 �1 �1
Pydub Mistral AI CrewAI Agents File Output Gmail API
- Input:
EarningsCall.wav(30-minute meeting recording) - Generated Files:
meeting_minutes/summary.txtmeeting_minutes/action_items.txtmeeting_minutes/sentiment.txt
- Email: Professional meeting minutes sent to stakeholders
- Transcription: ~2x audio length processing time
- Summarization: <30 seconds for typical meeting
- Email delivery: Real-time
Contributions, bug reports, and feature requests are warmly welcome! Feel free to open issues or submit pull requests.
This project is licensed under the MIT License.