ollama-mcp

spanchal001/ollama-mcp

★ 0 stars Python 🤖 AI/LLM Updated 5d ago

MCP server exposing local Ollama models to Claude Code for delegated work (summaries, extraction, code review/gen, sessions) with token-savings telemetry

View on GitHub →

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "ollama-mcp": {
      "command": "uvx",
      "args": [
        "ollama-mcp"
      ]
    }
  }
}

Or install with pip: pip install ollama-mcp

README Excerpt

MCP server that lets Claude Code delegate tasks to a local Ollama model (default `qwen3-coder:30b`). Ships 12 tools covering file Q&A, summarization, structured extraction, code review, multi-turn sessions, and bulk fan-out. The point: cheap/long/bulk work runs locally instead of burning Claude tokens, and sensitive code never leaves the machine.

Tools (7)

OLLAMA_HOSTOLLAMA_MCP_FILE_ROOTOLLAMA_MCP_MAX_FILE_BYTESOLLAMA_MCP_MAX_INFLIGHTOLLAMA_MCP_METRICS_DBOLLAMA_MCP_REQUEST_TIMEOUT_SECOLLAMA_MODEL

ollama-mcp

Quick Install

README Excerpt

Tools (7)

Related AI/LLM Servers