LM Studio Toolpack
Extends local LLMs with capabilities like web search, code execution, and persistent memory.
概要
This toolpack provides a versatile set of Model Context Protocol (MCP) tools specifically designed to enhance local Large Language Models. It enables LLMs to perform dynamic web searches using DuckDuckGo, execute Python code within a secure sandbox with access to mathematical libraries like NumPy and SymPy, and maintain long-term memory for more consistent and context-aware interactions. This collection empowers local LLMs with advanced agentic capabilities, making them more capable and useful for a variety of tasks.
主な機能
- Web Search (DuckDuckGo)
- Python Code Sandbox (Numpy, SymPy)
- Agent Long-term Memory
- 7 GitHub stars
ユースケース
- Enabling LLMs to solve complex mathematical problems
- Equipping LLM agents with persistent memory for continuous learning
- Providing LLMs with real-time information via web search