A Python-based automation tool that uses Llama 3 (via Ollama) to transform raw meeting transcripts into professional, structured meeting notes. This project is designed for users who need fast, accurate summaries without sending sensitive data to the cloud.
- Bulk Processing: Automatically detects and processes all
.txtfiles in the project directory. - Structured AI Insights: Uses custom prompt engineering to extract:
- Executive Summary: A concise overview of the discussion.
- Decision Tracking: Clearly listed agreements and conclusions.
- Action Items: Task-oriented lists with owners and deadlines.
- Privacy-First: Processes all data locally using Ollama. No data leaves your machine.
- Clean Output: Generates ready-to-use Markdown (.md) reports for documentation and sharing.
- Language: Python 3.x
- LLM Engine: Ollama
- Model: Llama 3 (8B)
- Libraries:
ollama,os
- Install Ollama
- Download the model:
ollama run llama3.2:1b # Or llama3 for higher accuracy
Clone this repository and install the required dependencies:
pip install ollama openai-whisper- Place your meeting transcript (.txt) or audio file (.mp3) in the project directory.
- Run the generator:
python app.py
- Find your structured notes in the generated
_summary.mdfile.
Input (Transcript):
John: We need to launch the beta by Friday. Sarah, can you finalize the UI?
Sarah: Yes, I'll have it done by Wednesday.
Mike: I will handle the server deployment.
AI-Generated Output:
The team discussed the upcoming beta launch timeline and assigned final preparation tasks.
- The Beta launch date is officially set for this Friday.
- Sarah: Finalize UI design (Deadline: Wednesday)
- Mike: Handle server deployment
This project was built with data security as a priority. By utilizing local LLMs via Ollama, sensitive meeting discussions remain on your local hardware, making it suitable for corporate environments with strict data privacy policies.