Ollama Claude

概要

This MCP server integrates your local Ollama instance with Claude Code, empowering Claude to offload resource-intensive coding tasks to your local models. It significantly reduces Anthropic API token consumption by performing code generation, explanation, review, and refactoring tasks on your local machine. Leveraging local compute resources, Ollama Claude works across any Claude Code project, offering both string-based and highly efficient file-aware tools, while Claude maintains oversight and refines the results for optimal output.

主な機能

• Delegates coding tasks to local Ollama models to reduce API token usage
• 0 GitHub stars
• Supports automatic orchestration, allowing Claude to intelligently delegate tasks to Ollama
• Offers a suite of string-based tools for common coding operations like generate, explain, review, and fix code
• Provides file-aware tools to process code directly from file paths, saving up to 98.75% in tokens
• Customizable default models and tool prompts for optimized local LLM performance

ユースケース

• Generate new code, functions, or unit tests using local models orchestrated by Claude
• Review, explain, or refactor existing code within files to dramatically cut cloud API token costs
• Analyze relationships and dependencies across multiple code files locally for deeper insights without exposing content to cloud APIs