🧠 MemoryLens

Local-first AI memory assistant for your meetings.
Record → Transcribe → Extract Tasks → Search — all on your own machine.

Download • Features • Architecture • Getting Started • API Docs • Roadmap • Contributing

---

📥 Download

Get the latest installer for Windows from the Releases page.

1. Download MemoryLens Setup 0.2.0.exe.
2. Run the installer and follow the instructions.
3. (Optional) Install the MemoryLens Browser Extension for speaker identification.

---

✨ Features

Feature	Description
🎙️ Smart Audio Capture	Records both your mic and system audio (Zoom, Meet, Teams) simultaneously via WASAPI loopback
📝 Local Transcription	Powered by Faster-Whisper — no data leaves your machine
👥 Speaker Identification	Maps speaker names from browser extension signals + voice anchoring
✅ Task Extraction	Automatically detects action items like "Aman, please finish the audit by Friday"
🔍 Natural Language Search	Ask questions like "What did Rahul promise last week?"
🧑‍🤝‍🧑 People Tracker	Track who you met, what was discussed, and tasks assigned to each person
🔒 100% Private	All processing runs locally — audio, transcription, and LLM inference on your own hardware
🖥️ Desktop App	Cross-platform Electron wrapper with a modern Next.js UI

---

🏗️ Architecture

Meeting Platform (Google Meet / Zoom / Teams)
       │
Browser Extension  → sends speaker events via HTTP
       │
┌──────────────────────────────────────────────────┐
│               MemoryLens Desktop Agent            │
│                                                   │
│   ┌─────────────┐  ┌──────────────┐  ┌─────────┐ │
│   │    Audio    │→ │ Transcription│→ │ Speaker │ │
│   │  Recorder   │  │  (Whisper)   │  │ Mapper  │ │
│   └─────────────┘  └──────────────┘  └─────────┘ │
│          │                │               │       │
│   ┌──────────────────────────────────────────────┐│
│   │               Memory Engine                  ││
│   │  ┌────────────┐ ┌───────────┐ ┌───────────┐  ││
│   │  │    Task    │ │  Summary  │ │  Meeting  │  ││
│   │  │ Extractor  │ │ Generator │ │ Processor │  ││
│   │  └────────────┘ └───────────┘ └───────────┘  ││
│   └──────────────────────────────────────────────┘│
│          │                                        │
│   ┌──────────────┐   ┌────────────────┐           │
│   │  SQLite DB   │   │ Semantic Search │           │
│   └──────────────┘   └────────────────┘           │
│          │                                        │
│   ┌──────────────┐                                │
│   │   FastAPI    │  ← REST API (localhost:8000)   │
│   └──────────────┘                                │
└──────────────────────────────────────────────────┘
          │
   ┌──────────────┐
   │  Electron +  │
   │  Next.js UI  │
   └──────────────┘

---

🚀 Getting Started

Prerequisites

• Python 3.10+
• Node.js 18+
• A working microphone (system audio capture available on Windows via WASAPI)

1. Clone the repository

git clone https://github.com/Yashkush06/MemoryLens.git
cd MemoryLens

2. Set up the backend

cd memorylens
pip install -r requirements.txt

3. Configure environment

# Copy the example env file and fill in your values
cp .env.example .env

Important: Set MEMORYLENS_SECRET_KEY to a strong random string before running. Generate one with: python -c "import secrets; print(secrets.token_hex(32))"

4. Run the backend

python scripts/run_agent.py

The API will be live at http://localhost:8000.
Open http://localhost:8000/docs for the interactive Swagger UI.

5. Run the frontend (Development)

cd frontend
npm install
npm run dev

The UI will be at http://localhost:3000.

6. Run the Desktop App (Electron)

cd electron
npm install
npm start

---

📡 API Docs

Method	Endpoint	Auth	Description
GET	/	—	Health check
POST	/auth/register	—	Register a new account
POST	/auth/login	—	Login and get JWT
GET	/auth/me	✅	Get current user
POST	/meeting/start	✅	Start recording a meeting
POST	/meeting/end	✅	Stop recording and process
GET	/meetings	✅	List all meetings
GET	/meeting/{id}	✅	Get meeting detail + transcript
POST	/upload_meeting	✅	Upload an existing audio file
GET	/tasks	✅	List extracted tasks
PATCH	/tasks/{id}	✅	Toggle task completion
GET	/people	✅	List known people
GET	/people/{id}/insights	✅	Get insights for a person
POST	/search	✅	Natural language search
GET	/system/models/status	—	Check AI model status
POST	/system/models/download	—	Download Whisper model

Full interactive docs at: http://localhost:8000/docs

---

⚙️ Configuration

All settings are controlled via environment variables prefixed with MEMORYLENS_.
Copy .env.example to .env to get started.

Variable	Default	Description
MEMORYLENS_SECRET_KEY	(required)	JWT signing secret — change this
MEMORYLENS_API_HOST	127.0.0.1	API bind address
MEMORYLENS_API_PORT	8000	API port
MEMORYLENS_WHISPER_MODEL_SIZE	base	Model size: tiny, base, small, medium, large-v2
MEMORYLENS_WHISPER_DEVICE	cpu	cpu or cuda
MEMORYLENS_USE_OLLAMA	false	Enable local Ollama LLM
MEMORYLENS_OLLAMA_MODEL	llama3	Ollama model to use
MEMORYLENS_LOG_LEVEL	INFO	Logging level

---

📁 Project Structure

MemoryLens/
├── memorylens/            # Python backend
│   ├── api/               # FastAPI server and routes
│   ├── audio/             # Audio recording (mic + WASAPI loopback)
│   ├── transcription/     # Faster-Whisper transcription engine
│   ├── speaker_mapping/   # Speaker identification from events
│   ├── memory_engine/     # Task extraction, summaries, processing
│   ├── search/            # Hybrid text + semantic search
│   ├── database/          # SQLAlchemy models and CRUD
│   ├── extension_api/     # Browser extension interface
│   ├── config/            # Pydantic settings
│   ├── shared/            # Logging, utilities
│   └── requirements.txt
├── frontend/              # Next.js UI
│   └── src/
│       ├── app/           # Next.js App Router pages
│       └── components/    # React components
├── electron/              # Desktop app wrapper
│   ├── main.js            # Electron main process
│   └── preload.js         # Electron preload
├── .env.example           # Environment configuration template
└── .gitignore

---

🗺️ Roadmap

• Audio capture (mic + system audio via WASAPI)
• Faster-Whisper transcription
• Task extraction from transcripts
• Natural language memory search
• JWT authentication
• Electron desktop app
• Real speaker diarization (pyannote.audio)
• LLM-powered task extraction via Ollama
• Semantic embeddings (sentence-transformers)
• Browser extension (Chrome/Firefox)
• Conversation replay with audio clips
• Export meetings to PDF / Markdown

---

🤝 Contributing

Contributions are welcome! Please read CONTRIBUTING.md to get started.

---

📄 License

This project is licensed under the MIT License — see LICENSE for details.

---

Made with ❤️ by Yash