Local LLM MCP Server

Local LLM MCP Server

A Model Context Protocol (MCP) server that bridges local LLMs running in LM Studio with Claude Desktop and other MCP clients. Keep your sensitive data private by running AI tasks locally while seamlessly integrating with cloud-based AI assistants.

🌟 Features

🔒 Privacy-First Design

• Local Processing: All sensitive data stays on your machine
• No Cloud Exposure: Private analysis, code review, and content processing happens locally
• Privacy Levels: Configurable privacy protection (strict, moderate, minimal)
• No Telemetry: Zero usage tracking or data collection

🤖 Dynamic Multi-Model Support

• Auto-Discovery: Automatically detects all models loaded in LM Studio
• Flexible Selection: Use different models for different tasks
• Runtime Switching: Change default models during your session
• Per-Request Override: Specify model for individual requests
• Smart Initialization: First available model auto-selected as default

🛠️ Comprehensive Tool Suite

Local Reasoning - General-purpose AI tasks with complete privacy

• Complex problem solving and multi-step reasoning
• Question answering and task planning
• Context-aware responses

Private Analysis - 7 analysis types for sensitive content

• Sentiment Analysis (domain-aware)
• Entity Extraction (people, orgs, locations, domain-specific)
• Content Classification
• Summarization with key points
• Privacy Scanning (PII, GDPR compliance)
• Security Auditing (vulnerabilities, misconfigurations)

Secure Rewriting - Transform text while maintaining privacy

• Style adaptation (formal, casual, professional)
• Sensitive information removal
• Privacy-preserving transformations

Code Analysis - Local code review and security

• Security vulnerability detection
• Code quality assessment
• Bug detection and optimization suggestions

Template Completion - Intelligent form and document filling

🎯 Domain-Specific Intelligence

Specialized analysis for:

• Medical: Healthcare context, HIPAA compliance, clinical terminology
• Legal: Legal terminology, regulatory compliance, confidentiality
• Financial: Financial regulations, market analysis, data protection
• Technical: Software development, engineering contexts
• Academic: Scholarly research, methodology, citations

🚀 Quick Start

Prerequisites

1. Node.js 18+

   node --version  # Should be >= 18.0.0
   

2. LM Studio

   • Download from lmstudio.ai
   • Load at least one model (e.g., Llama 3.2, Qwen, Mistral)
   • Start the local server (Server tab → Start Server)
   • Default URL: http://localhost:1234

Installation

git clone https://github.com/yourusername/local-llm-mcp-server.git
cd local-llm-mcp-server
npm install
npm run build

Configure Claude Desktop

Edit your Claude Desktop config file:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

Add this configuration:

{
  "mcpServers": {
    "local-llm": {
      "command": "node",
      "args": ["/absolute/path/to/local-llm-mcp-server/dist/index.js"]
    }
  }
}

Important: Use the absolute path to your installation.

Start Using

1. Restart Claude Desktop - The server starts automatically
2. Discover Models - Read resource local://models to see available models
3. Try It - Ask Claude to use the local_reasoning tool with a simple prompt

The server automatically:

• Discovers all models loaded in LM Studio
• Sets the first model as default
• Provides full capability documentation via local://capabilities

Convenience Scripts

For easier server management, use the included scripts:

# Start in local mode (stdio - for Claude Desktop)
npm run start:local

# Start in remote mode (HTTP - for network access)
npm run start:remote

# Start in secure mode (HTTPS - encrypted network access)
npm run start:https

# Start in dual mode (stdio + HTTP for both local and remote)
npm run start:dual

# Generate SSL certificates for HTTPS
npm run generate:certs

# Stop all running servers
npm run stop

See SCRIPTS_GUIDE.md for detailed usage.

🌐 Remote Network Access

Access the server from other devices on your home network or connect Claude Desktop remotely!

Connect Claude Desktop Remotely

Quick Start (3 steps):

# 1. Start HTTPS server
npm run start:https

# 2. Add to claude_desktop_config.json:
{
  "mcpServers": {
    "local-llm-remote": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://localhost:301