# 8b-is/smart-tree [Health: Active]

**Category:** 📂 File Systems  
**Repository:** https://github.com/8b-is/smart-tree  
**GitHub Stars:** 267  
**Views:** 6  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/8b-is-smart-tree

## Description
AI-native directory visualization with semantic analysis, ultra-compressed formats for AI consumption, and 10x token reduction. Supports quantum-semantic mode with intelligent file categorization.

## Claude Desktop Quick Installation
Heuristic fallback — verify the package name and runner against the repository README before running it. Uses `npx` (confidence: low):

```json
"mcpServers": {
  "smart-tree": {
    "command": "npx",
    "args": ["-y","8b-is-smart-tree"]
  }
}
```

## Documentation

## What 8b-is/smart-tree MCP server does

8b-is/smart-tree MCP server connects an AI client to a local directory and exposes Smart Tree’s repository-inspection functions through Model Context Protocol. Its core use is turning a directory into an AI-oriented representation rather than returning only a conventional tree listing. The project description also identifies semantic analysis, intelligent file categorization, and compressed output intended to reduce the amount of context sent to an agent.

The surrounding Smart Tree CLI supports several related workflows. Agents can request a quick tree view, search file contents, save session information, and create or retrieve collaborative memories. The README also documents an `ask_user` MCP tool that can pause an agent workflow and request input through the dashboard.

## How it works

Smart Tree is implemented as a local Rust application. The README lists a Model Context Protocol option that communicates over stdio, so the MCP client starts the local program rather than connecting to a hosted endpoint. A client can therefore inspect files available to the machine running Smart Tree, subject to that process’s filesystem access.

Output can be selected for AI use with the CLI’s AI mode and compression flags. The documented quantum mode is intended to produce an even smaller representation, and the project description associates that mode with intelligent categorization. These formats are useful when an agent needs repository structure or selected content while keeping prompt size down.

The project also includes a client menu with entries for a basic tree, an interactive file browser, a terminal interface, a web dashboard, an HTTP daemon, and the stdio MCP server. These are additional application modes rather than separate MCP services.

## Setup and configuration

The project can be installed on Linux, macOS, or WSL with its shell installer or through `cargo install smart-tree`. Windows has a PowerShell installer, and the README also documents a Homebrew formula for macOS. After installation, the executable is named `st`.

Claude Desktop configuration can be handled from the CLI with `st --mcp-install`; the README says this creates a backup and preserves existing configuration. `st --mcp-status` checks the installation state, while `st --mcp-uninstall` removes the configuration. The repository also includes interactive setup and management scripts for MCP setup, health checks, and client-mode selection.

## Tools and capabilities

The documented capabilities include:

- Directory visualization for conventional and AI-focused views
- Compressed output for reducing codebase context size
- Quantum mode with more aggressive compression
- Fast content search, including searches such as `TODO`
- Collaborative memory lookup and session-memory anchoring
- An `ask_user` MCP interaction for requesting human input
- Interactive browsing through the Spicy TUI
- Security scans for documented MCP supply-chain attack patterns

8b-is/smart-tree MCP server is therefore most relevant to local repository exploration and agent context preparation. Its memory, dashboard, and security features broaden the CLI, but they should not be treated as requirements for using the basic MCP mode.

_Full upstream README: https://allmcps.com/mcp/8b-is-smart-tree/readme_