Local MCP Server

Run a full Model Context Protocol (MCP) server in Termux. Exposes Read, Bash, Grep, glob tools for local Ollama models. Privacy-first, offline-capable, ~10MB.

Local MCP Server for Termux

Run a full MCP server in Termux that exposes the same tools as Amp but for local AI models via Ollama.

What Is MCP?

The Model Context Protocol is the standard for AI tool integration. This server implements it locally, so your Ollama models can:

• Read files and directories
• Execute bash commands
• Search code with grep
• Pattern match files with glob

Perfect For

• Developers who want local AI without cloud dependency
• Privacy-conscious users keeping code on-device
• Termux power users building AI workflows
• Offline development environments

What's Included

• Full MCP server implementation (Node.js, ~10MB)
• Tool wrappers for local CLI utilities
• Path sandboxing for security
• HTTP and stdio transport options
• Amp-compatible interface (easy switching)
• Health checks and graceful shutdown

Integration Examples

• Connect Ollama models to local file system
• Use with Claude Desktop as local tool provider
• Build custom AI workflows with tool access

Quick Start

bash

# Install dependencies
pnpm install

# Setup server
bash scripts/setup.sh

# Start MCP server
mcp-start

# Verify health
mcp-health

# Connect Ollama model
ollama run llama3.2 --mcp-server http://localhost:8080

Configuration

json

{
  "server": {
    "port": 8080,
    "transport": "http",
    "sandbox": "~/.openclaw/mcp-sandbox"
  },
  "tools": {
    "read": true,
    "bash": true,
    "grep": true,
    "glob": true
  },
  "security": {
    "allowedPaths": ["~/projects", "~/.openclaw"],
    "forbiddenCommands": ["rm -rf", "dd", "mkfs"]
  }
}

Requirements

• Termux (Android) or Linux environment
• Node.js v16+
• Ollama installed (optional, for LLM integration)
• pnpm or npm

Support

Email: [e-mail]