AI Meeting Intelligence Platform is a full‑stack system that converts meeting recordings into structured, searchable, and insightful summaries.
It combines automatic transcription, semantic analysis, sentiment detection, and AI summarization to give teams an instant overview of discussions, decisions, and action items.
Built with FastAPI, React, and Ollama, it integrates whisper.cpp (or faster‑whisper) for transcription and ChromaDB for vector‑based semantic search — providing an efficient, privacy‑preserving, and self‑contained solution for meeting intelligence.
- Accurate Transcription (whisper.cpp / faster‑whisper)
- Topic Extraction & Summaries (LLM via Ollama)
- Sentiment Insights per segment
- Semantic Search over transcripts (ChromaDB)
- Action Items & Decisions extraction
- Fully Local / Offline‑friendly (no external data sharing)
- Simple Web UI for upload, processing, and review
Frontend (React + Tailwind + Vite)
│
▼
Backend (FastAPI + SQLite + ChromaDB)
│
├── Transcription (whisper.cpp / faster-whisper)
├── Embeddings (Ollama)
├── Semantic Indexing (Chroma)
└── LLM Summarization + Sentiment + Topics
| Directory | Description |
|---|---|
| backend/ | FastAPI backend with SQLite DB, ChromaDB, whisper.cpp, and Ollama integrations. |
| frontend/ | React + Tailwind + Vite interface for uploading meetings and visualizing insights. |
| docs/ | Architecture and API documentation. |
| docs/tech-writeup/ | Technical write‑ups: architecture, pipeline, and challenges (README.md, api.md, architecture.md). |
cd backend
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env- Update
.envwith:WHISPER_BIN→ path to whisper.cpp binaryWHISPER_MODEL→ path to your model file
- Start Ollama and pull required models:
ollama pull llama3 ollama pull nomic-embed-text
- Run the backend:
uvicorn app.main:app --reload
cd frontend
npm install
npm run devOpen: http://localhost:5173
- Create a meeting and upload audio/video.
- Click Process to run transcription → embeddings → indexing → LLM insights.
- Explore: summary, topics, sentiments, transcript, and search highlights.
| Engine | Description | Setup |
|---|---|---|
| whisper.cpp | Local, fast transcription with low resource usage. | Configure WHISPER_BIN and WHISPER_MODEL in .env. |
| faster‑whisper | Python engine; supports optional diarization. | Set TRANSCRIPTION_ENGINE=faster_whisper; diarization via pyannote requires HF_TOKEN. |
- ChromaDB:
backend/data/chroma - SQLite:
backend/data/app.db - Uploads:
backend/data/uploads
docs/tech-writeup/README.md— Technical overview (architecture, pipeline, challenges)docs/tech-writeup/api.md— API endpoints and usagedocs/tech-writeup/architecture.md— System design details & diagrams
- ✅ Working backend & frontend codebase
- ✅ API, docs, and setup instructions
- ✅ Post‑meeting pipeline (transcription → insights → search)
Released under the MIT License. See LICENSE for details.