Prompt Cleaner

关于

Prompt Cleaner is a TypeScript-based Model Context Protocol (MCP) server designed to act as a crucial intermediary for large language model (LLM) interactions. It provides a robust `cleaner` tool that processes raw prompts, applying secret redaction, ensuring structured output via JSON schemas, and normalizing content for client compatibility. This server enhances security by scrubbing sensitive patterns and improves reliability by standardizing LLM responses, offering a configurable and deterministic interface for various OpenAI-compatible LLM APIs.

主要功能

• Configurable LLM integration (API base, key, model, timeout)
• Automatic secret redaction from logs and outputs
• Deterministic single-model policy for consistent behavior
• Normalizes LLM output (e.g., JSON to plain text) for client compatibility
• 1 GitHub stars
• Cleans and structures raw prompts for LLMs

使用案例

• Standardizing LLM outputs for diverse client applications and agents
• Preprocessing user prompts before sending them to an LLM
• Ensuring sensitive data is removed from LLM inputs and responses