A Rust-based TUI application that helps you discover and understand your Homebrew-installed tools. It indexes local man/help pages into a vector store and uses a local Ollama model to recommend tools and usage examples based on your query.

• 📦 Package Discovery: Automatically lists all Homebrew-installed packages
• 🤖 AI-Powered Recommendations: Uses local Ollama to suggest tools based on your needs
• 💡 Usage Examples: Provides practical command-line examples for recommended tools
• ⌨️ Interactive TUI: Clean, intuitive terminal interface built with Ratatui

Before running this application, ensure you have:

1. Rust (1.70 or later)

   curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

2. Homebrew (macOS package manager)

   /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

3. Ollama (local LLM runtime)

   brew install ollama

4. Ollama Models

   ollama pull all-minilm
   ollama pull qwen3-coder:480b-cloud

1. Clone the repository and enter the directory:

   git clone <repo_url>
   cd mac-aid

2. Build the project:

   cargo build --release

3. Run the application:

   cargo run --release

• Enter: Submit your query
• Esc: Clear input
• q: Quit
• Ctrl + r: Rebuild knowledge base
• Shift + R: Reload index data
• ↑/↓: Scroll response

Run a one-shot query directly from the terminal (no TUI, no RAG retrieval):

mac-aid how to compress file
mac-aid "convert image format"

This uses local Ollama with the configured generation model and prints the answer to stdout.

Try asking questions like:

• "I need to process JSON files"
• "How can I convert images to different formats?"
• "Show me tools for monitoring network traffic"
• "I want to edit videos from the command line"
• "What tools can help me work with Docker?"

┌──────────────────────────────────────────────┐
│  Enter your need                           │
│ [type your query here]                       │
└──────────────────────────────────────────────┘
┌──────────────────────────────────────────────┐
│ Status: Ready / Indexing progress messages   │
└──────────────────────────────────────────────┘
┌──────────────────────────────────────────────┐
│ 💡 Recommendation                            │
│ AI suggestions and usage examples appear here│
└──────────────────────────────────────────────┘

The interactive query model is set in src/main.rs and defaults to qwen3-coder:480b-cloud:

let ollama = OllamaClient::new("qwen3-coder:480b-cloud".to_string());

Pull and set any model you prefer by updating this line.

Models are configured in ~/.mac-aid/config.json and auto-created on first run:

{
  "ollama_model": "qwen3-coder:480b-cloud",
  "embedding_model": "all-minilm"
}

If your Ollama instance is running on a different host/port, edit src/ollama.rs:

impl OllamaClient {
    pub fn new(model: String) -> Self {
        Self {
            client: Client::new(),
            base_url: "http://your-host:port".to_string(),
            model,
        }
    }
}

Make sure Homebrew is installed and accessible in your PATH:

which brew
brew --version

1. Ensure Ollama is running:

   ollama serve

2. Verify the model is available:

   ollama list

3. Test the API manually:

curl http://localhost:11434/api/generate -d '{ "model": "qwen3-coder:480b-cloud", "prompt": "Hello", "stream": false }'

### Non-blocking rebuild/reload

- Rebuild and reload run in the background; the TUI stays responsive.
- Queries made during rebuild/reload skip RAG retrieval and directly use Ollama.

## Development

### Project Structure

src/ ├── main.rs ├── app.rs ├── ui.rs ├── brew.rs ├── indexer.rs ├── vector_store.rs ├── rag.rs ├── langchain_integration.rs ├── kb_builder.rs └── log.rs

Additional docs:

• Chinese README: README.zh-CN.md

### Building for Development

```bash
cargo build
cargo run

cargo test

• Database: ~/.mac-aid/commands.db
• Logs: ~/.mac-aid/error.log, ~/.mac-aid/info.log

MIT

Contributions are welcome! Please feel free to submit issues or pull requests.