Interactive Physics Assistant (IPA)

The widespread adoption of Large Language Models (LLMs) in education has led to concerning trends where students substitute AI assistance for critical thinking. This work presents an intelligent physics tutoring system designed to leverage LLMs while preserving essential problem-solving skills. Our system integrates open-source LLMs with Model Context Protocol (MCP) servers, enabling access to specialized physics computation tools and visualizations beyond standard LLM capabilities. Our system guides students through structured problem-solving processes. This methodology aims to strengthen analytical thinking while making quality physics education more accessible. The system successfully promotes deeper conceptual understanding while maintaining student agency in the learning process.