trending
•

7h ago

PIIGhost - Keep personal data out of your LLM prompts

Send a prompt to GPT, Claude or Gemini and the provider sees everything in it, including your users' personal data. piighost swaps that data for placeholders like <> before the model sees it, then restores the real values in the reply. Your tools still receive the real values. Works on text, conversations and agents, with LangChain, Pydantic AI and LlamaIndex connectors and an OpenAI-compatible proxy. MIT, open source.