smart-tree

Smart Tree is a fast, AI-friendly MCP server built around a Rust-native directory visualization tool. When run in MCP mode, it exposes a protocol-enabled interface that allows clients to interact with a live file system view, perform quick-tree queries, and coordinate with an integrated AI assistant for memory-aware context management. The server ships with a rich set of features accessible through the CLI and MCP-driven hooks, including a Spicy TUI mode, a web dashboard, an HTTP daemon, and memory-conscious tooling to help you build smarter AI-assisted workflows. To use it, start the MCP server (for example via st --mcp) and connect clients that support the MCP protocol; you’ll be able to browse directories, request structured tree representations, visualize activity with Wave Compass, and leverage the built-in security scanner when scanning for potentially malicious MCP integrations. The toolset also includes menu-driven access to different modes like Basic Tree View, Spicy TUI, Terminal Interface, Web Dashboard, and an MCP-stdio server for seamless integration with Claude/AI assistants.

Once running, you can access the MCP-enabled features through the client or CLI: the server can present a persistent directory visualization, respond to quick_tree-style queries, and provide hooks-management for Claude Code-based integrations. It’s designed to be lightweight and fast, with modern memory-preserving options (Aye Consciousness) and compression optimizations to keep your context window small yet informative. Use the MCP dashboard and health checks to monitor activity, and consider enabling the security scanner to detect supply-chain risks in connected MCP servers.

Prerequisites:

• A supported OS (Linux/macOS/Windows with appropriate tooling)
• Rust tooling or a pre-built binary (as provided by the project)
• Network access for installation scripts or package sources

Option A — One-line installer (recommended for quick setup):

1. Run the one-line installer to fetch and install Smart Tree:

curl -sSL https://raw.githubusercontent.com/8b-is/smart-tree/main/scripts/install.sh | bash

2. Verify installation by checking the version:

st --version

3. Start MCP mode:

st --mcp

Option B — Install via cargo (source build):

1. Ensure Rust and cargo are installed on the system.
2. Install directly from GitHub:

cargo install --git https://github.com/8b-is/smart-tree

3. Confirm the binary is in your PATH and run MCP mode:

st --mcp

Tips:

• After installation, you can run st --cleanup to scan for and remove malicious AI integrations.
• You can also use the developer/menu scripts for setup and testing: ./scripts/interactive_setup.sh, ./scripts/manage.sh menu, ./scripts/manage.sh client.

Tips and common considerations:

• Security scanning: use st --cleanup and st --security-scan . to detect and mitigate MCP-related risks.
• If you install new MCP servers, re-run the cleanup to ensure you’re not introducing unwanted hooks or external network calls.
• The MCP mode is designed to be quiet (no extraneous stderr output) for a cleaner protocol surface; enable verbose logging only when debugging.
• For best results with AI assistants, leverage features like Aye Consciousness, Memory Anchoring, and quick_tree to keep context concise and useful.
• If you encounter issues, verify that your environment variables are set as expected by the server and that port access for the dashboard and MCP endpoints is allowed by your firewall.