This project implements a multi-agent pipeline that preprocesses, simplifies, and reviews user prompts before sending them to an LLM—removing redundant detail and sensitive information where possible (data minimization for Trust & Responsible AI). The app mimics a ChatGPT-style experience with configurable privacy levels (low / medium / high). A simplification agent chain produces a minimized prompt and a final answer for the user.
@khoidm2004
owner
