Local LLMs MCP servers & tools

8 Model Context Protocol servers tagged “Local LLMs”, ranked by usage. Connect any of them to Claude Desktop, Cursor or Windsurf in one step.

What qualifies: Its tools work with locally run models (Ollama, LM Studio, llama.cpp) (its API, an account in it, or its data). Supporting any one of the products named is enough. Each server was checked against this, not tagged by keyword.

Search “Local LLMs” with filters

TokenMark
Measured local-LLM speeds and hardware-aware model picks for Strix Halo, DGX Spark and Mac.
Conversational AI
Upvotes: 0Installs: 0GitHub stars: —
~alter mcp Ollama
Wraps local Ollama models so bulk generation runs on your own hardware, not a paid API.
Conversational AI
Upvotes: 0Installs: 0GitHub stars: —
Openclaude MCP
FastMCP 3.2 server plus Webapp for using the leaked Claude Code locally with ollama and open weight models. Kairos and cat included
Conversational AI
Upvotes: 0Installs: 0GitHub stars: 7
MCP Server Ollama Bridge
jaspertvdm
Bridge to local Ollama LLM server. Run Llama, Mistral, Qwen and other local models through MCP.
Conversational AI
Upvotes: 0Installs: 0GitHub stars: 4
Codebase Agent MCP
FI-Mihej
Delegates documentation and code analysis, including calls to other MCP servers, to a local or low-cost OpenAI-compatible LLM to save top-tier model tokens.
Coding Agents
Upvotes: 0Installs: 0GitHub stars: 3
Local Llm Worker
Local LLM does Claude's bulk work: reads logs and files, researches the web, writes test-gated code.
Coding Agents
Upvotes: 0Installs: 0GitHub stars: 2
AI Guardian
Governed local-LLM (Ollama) observability: model policy, prompt scanner, 18 tools.
Monitoring
Upvotes: 0Installs: 0GitHub stars: —
Code Firewall MCP
Structural similarity-based code filter. Stops malicious code pattern reaching execution tools.
Security
Upvotes: 0Installs: 0GitHub stars: —