Show HN: Aiproxy – Redact PII locally before it reaches Claude/OpenAI (github.com)

Aiproxy has been launched as a free anonymizing reverse proxy designed to protect sensitive data before it interacts with large language models (LLMs) such as Claude and OpenAI's offerings. By scrubbing personally identifiable information (PII) like names and other confidential content using a local LLM, aiproxy ensures that user data remains private. The tool scrubs PII before sending it to the upstream AI providers and restores the originals in the response, maintaining a secure and reversible anonymization process stored locally in an encrypted database. This project is significant for the AI/ML community because it addresses growing concerns about data privacy and compliance with regulations like GDPR. Aiproxy offers a simple, plug-and-play architecture for filtering content, handling various data types including text, images, and PDFs. Beyond its immediate privacy benefits, it allows for customization through a pluggable filter interface and scalable organizational deployment options. By enabling users to leverage AI tools without compromising data security, aiproxy empowers developers and enterprises to adopt AI responsibly, ensuring they uphold data protection standards while benefiting from advanced machine learning applications.