Tillu - Your Personal AI Assistant

"Your AI companion that truly knows you"

Tillu is a 100% local, privacy-first AI assistant that permanently remembers everything about you. Built for anyone who values privacy and wants a truly personalized AI experience without compromising their data.

🎯 What Tillu Does

• Remembers Everything: Your resume, skills, contact information, preferences - stored locally forever
• Intelligent Conversations: Answers questions using YOUR personal context and information
• Personalized Responses: Generates content tailored to your experience and writing style
• 100% Private: All data stays on your machine, AI runs locally with no cloud dependency
• Semantic Search: Understands meaning, not just keywords - finds relevant information intelligently

✨ Key Features

🔒 Privacy First

• Zero Cloud Dependency: Everything runs on your local machine
• No Data Transmission: Your information never leaves your computer
• Open Source: Audit the code yourself - complete transparency
• You Own Your Data: Full control over your personal information

🧠 Intelligent Context

• Semantic Understanding: Finds relevant information based on meaning
• RAG (Retrieval-Augmented Generation): AI responses grounded in your actual data
• Vector Search: Fast, accurate retrieval of contextual information
• Persistent Memory: Never repeat yourself - Tillu remembers

⚡ Performance

• Fast Responses: < 3 seconds for most queries
• Efficient Search: < 100ms semantic search
• Lightweight: Runs smoothly on modern hardware
• Local Processing: No network latency

🚀 Quick Start

Installation

# 1. Clone the repository
git clone git@github.com:Manmit124/tillu.git
cd tillu

# 2. Install Ollama (Local AI Runtime)
# Download from https://ollama.ai and install
# Or use the provided script:
./scripts/install-ollama.sh

# 3. Download AI model
ollama pull llama3.2

# 4. Install backend dependencies
cd backend
npm install

# 5. Set up your personal data
# Edit files in data/personal/ with your information:
# - basic.json (name, contact, location)
# - professional.json (LinkedIn, GitHub, portfolio)
# - preferences.json (job preferences, interests)

# 6. Ingest your data into the vector database
npm run ingest

# 7. Start the backend server
npm run dev

📚 How It Works

Architecture

┌─────────────────────────────────────────────────────┐
│                   Your Computer                      │
│                                                      │
│  ┌──────────────┐      ┌─────────────────────────┐ │
│  │   Frontend   │◄────►│   Backend Server        │ │
│  │  (Coming     │      │   (Node.js/Express)     │ │
│  │   Soon)      │      │                         │ │
│  └──────────────┘      │  ┌──────────────────┐   │ │
│                        │  │  RAG Pipeline    │   │ │
│                        │  │  - Semantic      │   │ │
│                        │  │    Search        │   │ │
│                        │  │  - Context       │   │ │
│                        │  │    Building      │   │ │
│                        │  │  - Response      │   │ │
│                        │  │    Generation    │   │ │
│                        │  └──────────────────┘   │ │
│                        │                         │ │
│                        │  ┌──────────────────┐   │ │
│                        │  │  Vector Database │   │ │
│                        │  │  (SQLite + vss)  │   │ │
│                        │  └──────────────────┘   │ │
│                        └─────────────────────────┘ │
│                                 │                   │
│                                 ▼                   │
│                        ┌─────────────────────────┐ │
│                        │   Ollama (Local AI)     │ │
│                        │   - Llama 3.2 Model     │ │
│                        │   - Embeddings          │ │
│                        └─────────────────────────┘ │
│                                                      │
│  All processing happens locally   │
└─────────────────────────────────────────────────────┘

Technology Stack

Backend

• Node.js + Express: Fast, lightweight API server
• TypeScript: Type-safe development
• SQLite + sqlite-vss: Vector similarity search
• Transformers.js: Local embeddings (all-MiniLM-L6-v2)
• Ollama: Local AI model runtime

AI Layer

• Llama 3.2 (3B): Fast, efficient language model
• RAG Pipeline: Retrieval-Augmented Generation
• Semantic Search: 384-dimensional vector embeddings
• Context Building: Intelligent information retrieval

Frontend (Coming Soon)

• Chrome Extension for browser integration
• React + TypeScript for UI
• TailwindCSS for styling

🎯 Use Cases

Personal Knowledge Base

• Store and retrieve your personal information instantly
• Never forget important details about yourself
• Quick access to your resume, skills, and experience

Job Applications (Planned)

• Auto-fill job application forms
• Generate personalized cover letters
• Answer job-specific questions using your experience

Personal Assistant

• Answer questions about yourself
• Maintain context across conversations
• Learn and adapt to your preferences

📖 Documentation

• Setup Guide - Detailed installation instructions
• Implementation Plan - Complete development roadmap
• Vector Storage - How semantic search works
• API Documentation - API endpoints and usage

🔐 Privacy & Security

What We Collect

Nothing. Tillu doesn't collect, transmit, or store any data externally. All your personal information stays on your machine.

Where Your Data Lives

• Local SQLite Database: data/buddy.db
• Personal Files: data/personal/*.json
• Vector Embeddings: Generated and stored locally

Internet Usage

• Setup Only: Internet required to download AI models (~2GB) and dependencies
• After Setup: All processing happens locally - your data and conversations stay on your machine
• No Cloud Services: Unlike ChatGPT or other cloud AI, your information never gets sent to external servers

Local Processing

After initial setup (which requires internet to download models), all AI processing happens on your machine. Your conversations and personal data never leave your computer.

Open Source

Audit the entire codebase. No hidden telemetry, no analytics, no tracking.

🚧 Current Status

✅ Completed (Phase 1 & 2)

• Backend server with Express
• Ollama integration
• Vector database (SQLite + vss)
• Local embeddings (Transformers.js)
• Semantic search
• RAG pipeline
• Data ingestion
• API endpoints
• Documentation

🚀 Coming Soon (Phase 3+)

• Chrome extension
• Browser integration
• Form auto-fill
• Cover letter generation
• Multi-resume support
• Job tracking
• UI improvements

🤝 Contributing

Contributions are welcome! This is an open-source project built for privacy-conscious individuals.

How to Contribute

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📝 License

MIT License - See LICENSE for details.

You are free to use, modify, and distribute this software. No attribution required, though appreciated!

🙏 Acknowledgments

Built with:

• Ollama - Local AI runtime
• Transformers.js - Local embeddings
• sqlite-vss - Vector similarity search
• LangChain - RAG framework inspiration

💬 Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions

🌟 Why Tillu?

In a world where AI assistants send your data to the cloud, Tillu takes a different approach:

• Your data is yours - No company has access to your personal information
• Privacy by design - Built from the ground up with privacy as the core principle
• Truly personalized - AI that knows YOU, not generic responses
• Open and transparent - Audit every line of code
• No subscription - Free forever, runs on your hardware

---

Built with ❤️ for privacy-conscious individuals who want a truly personal AI assistant.

Note: Tillu is currently in active development. Phase 1 & 2 are complete. Chrome extension and advanced features coming soon!