Show HN: TSZ – Open-source guardrails and data security layer for LLM pipelines (github.com)

Thyris.AI has launched TSZ (Thyris Safe Zone), an open-source PII Detection and Guardrails System designed to provide a zero-trust layer for large language model (LLM) pipelines. This system is significant for the AI/ML community as it addresses critical data security concerns, enabling safe integration of LLMs and third-party APIs without risking the exposure of sensitive information like Personally Identifiable Information (PII). TSZ's key features include real-time PII detection and redaction using context-preserving placeholders, configurable guardrails, and management options for allowlists and blocklists. The system is implemented in Go with high performance, utilizing Redis caching for rapid data processing. It also allows for dynamic rule updates through APIs, enhancing adaptability in various application environments. TSZ provides comprehensive documentation, including quick start guides and API references for developers, along with official client libraries for both Go and Python. This open-source initiative not only fosters community contributions but also establishes a framework for securing LLM utilization, making it a pivotal advancement in ensuring data safety in AI applications.