PrivAiTe

Overview

Self-hosted PII redaction proxy for LLM APIs and agent CLIs. Scrubs names, emails and secrets from Claude Code, Codex and any OpenAI-compatible app before requests leave your machine, tool-call arguments included, then restores them in the reply. Local detection, zero telemetry.

Best for

Best for
Python developers who want local PII redaction for LLM and agent workflows without sending data to third-party services

Use cases

• Redact PII from LLM API requests before they leave the machine
• Protect agent CLI tool-call arguments for Claude Code and Codex
• Restore redacted entities in responses from OpenAI-compatible apps

Notes

Self-hosted PII redaction proxy for LLM APIs and agent CLIs. Scrubs names, emails and secrets from Claude Code, Codex and any OpenAI-compatible app before requests leave your machine, tool-call arguments included, then restores them in the reply. Local detection, zero telemetry.

47 stars on GitHub. Last updated 2026-10-05. Licensed BSD-3-Clause.

Use cases

• Redact PII from LLM API requests before they leave the machine
• Protect agent CLI tool-call arguments for Claude Code and Codex
• Restore redacted entities in responses from OpenAI-compatible apps

Pros

• Local detection with zero telemetry keeps data on your machine
• Restores redacted values in replies, so downstream workflows keep working
• Works across standard OpenAI-compatible APIs and popular agent CLIs

Cons

• Requires self-hosting and a local proxy setup to intercept traffic
• Small community with 47 GitHub stars and limited ecosystem traction
• Only covers OpenAI-compatible APIs and the specifically listed agent CLIs

Indexed from awesome-llmops and enriched against its public facts.