PrivAiTe: Self-hosted proxy that redacts PII from LLM calls, incl. tool-calls
PrivAiTe is a self-hosted proxy designed to redact Personally Identifiable Information (PII) from LLM API calls, including tool calls, before they reach external AI providers. It aims to enhance privacy by allowing users to keep sensitive data local while still leveraging cloud-based language models.
Background
- PrivAiTe is an open-source, self-hosted proxy that intercepts API calls to LLM providers (OpenAI, Anthropic, etc.) and automatically redacts personally identifiable information (PII) — names, emails, phone numbers, addresses — before the data reaches the LLM. It then restores the original text in responses.
- Why it matters: Companies want to use powerful third-party LLMs but risk leaking sensitive user data to the provider's servers. PrivAiTe lets them use those LLMs while keeping PII out of provider logs and training data — especially important for GDPR/HIPAA compliance.
- It uniquely handles tool-calls (function calling), a common LLM agent pattern that most existing redaction proxies don't support, making it relevant for developers building AI agents.
- Being self-hosted means no data is sent to another redaction service; all processing happens locally on the user's own infrastructure.