jotx

Your digital memory. Search everything you've copied or typed, instantly.

jot is a fully local, privacy-first AI assistant that remembers your clipboard history and terminal commands. Ask questions in natural language and get instant answers, no scrolling, no searching, just results.

✨ Features

• 🔍 Natural Language Search - Find things by meaning, not just keywords
• 🔒 100% Private - Everything stays on your machine, encrypted
• ⚡ Lightning Fast - Search 10,000+ items in milliseconds
• 🎨 Dual Interface - Beautiful GUI or blazing-fast CLI
• 🧠 Context Aware - Understands what you're looking for
• 🧩 Plugin Ready - Extend Jot with custom Rhai plugins that tap into lifecycle events.

🎯 Use Cases

Uses jotx ask (ja for short) for natural language search and search (js for short) for normal keyword search

# Instead of scrolling through terminal history
jotx ask "ssh command for staging server"
→ ssh [email protected] -i ~/.ssh/key.pem

# Find that email you copied hours ago (-c means search in clipboard history)
ja -c "email address from this morning"
→ [email protected]

# Find that yarn command to run the server
js "yarn"
→ yarn start

🚀 Quick Start

macOS & Linux

Copy and paste this into your terminal:

curl -fsSL https://raw.githubusercontent.com/Jeffawe/Jot/main/install.sh | bash

That's it! The installer will:

• ✅ Download the right binary for your system
• ✅ Install Ollama (local AI)
• ✅ Set up shell hooks
• ✅ Start the daemon

Alternative with wget:

wget -qO- https://raw.githubusercontent.com/Jeffawe/Jot/main/install.sh | bash

First Run

# Start monitoring (runs in background)
jotx run

# Search your history
jotx search "ssh"

# Ask questions
jotx ask "what was that git command from yesterday?"

📖 How It Works

Natural Language Search

1. Monitors clipboard and terminal commands using rust copypasta and shell hooks
2. Stores everything locally in a local SQLite database (~/.jotx/jotx.db)
3. Indexes content using embedding models for semantic search
4. Searches using natural language and pluggable LLm models (via ollama) to query db and give results fast

GUI Mode

Install the GUI version from https://github.com/Jeffawe/Jot/releases and look for the desktop release

⚙️ Configuration

Configuration file: ~/.jotx/config.toml

[llm]
provider = "ollama"
api_base = "http://localhost:11434"
model = "qwen2.5:3b"
max_tokens = 500
temperature = 0.7
max_history_results = 10

[search]
similarity_threshold = 0.5
max_results = 10
fuzzy_matching = true

[storage]
maintenance_interval_days = 7

🔒 Privacy & Security

jot is built privacy-first:

• ✅ 100% Local - No data ever leaves your machine
• ✅ No Telemetry - Zero analytics or tracking
• ✅ Configurable Exclusions - Block apps, files, or patterns (run jotx privacy)
• ✅ Open Source - Fully auditable code
• ✅ Clean Data - Clean data stored anytime easily (run jotx clean-data)

🛠️ Tech Stack

• Language: Rust 🦀
• Storage: SQLite
• Search: fastembed (embedding models)
• AI: Ollama
• GUI: Tauri (Rust + Web)
• CLI: clap for argument parsing

🤝 Contributing

Contributions are welcome! This project is built for learning Rust, so beginner-friendly PRs are encouraged.

Development Setup

# Clone the repo
git clone https://github.com/jeffawe/jot.git
cd jot

# Install dependencies (Full Setup)
make setup 

#Install dependencies (Run Rust code)
cargo build

# Run tests
cargo test

Areas for Contribution

• 🐛 Bug fixes
• 📝 AI development
• ✨ New search algorithms
• 🎨 UI/UX enhancements
• 🔧 Performance optimizations
• 🧪 Test coverage

📄 License

Apache License - see LICENSE for details

🙏 Acknowledgments

• Built with Rust
• Embeddings via fastembed
• AI via Ollama for running local LLMs
• Inspired by the need to remember things better

💬 Support

• 📫 Issues: GitHub Issues
• 💭 Discussions: GitHub Discussions
• 🐦 Twitter: @awagu_jeffery

---

Remember: Your digital memory, always at your fingertips. Never scroll through history again.

Built with ❤️ and Rust 🦀