All activity
Tyler Youngleft a comment
built Guardrail Layer after watching how easy it was for “harmless” LLM prompts to accidentally expose real production data. Traditional RBAC works at the app layer, but breaks down once you let users chat directly with databases. The goal was simple: make AI data access safe by default, not something you clean up after the model responds. Happy to answer questions about architecture, security...
Guardrail LayerRole-aware access control for AI talking to production data
Guardrail Layer is a security layer that sits between LLMs and real production data. It enforces role-aware field access, query-time filtering, and audit logging before data ever reaches the model. Unlike tools that clean responses after the fact, Guardrail Layer prevents unsafe queries entirely—making it possible to chat with live databases without leaking PII or internal-only data.
Guardrail LayerRole-aware access control for AI talking to production data
Tyler Youngleft a comment
Hey everyone! 👋 I built StampPDF after getting frustrated with bloated, cloud-only PDF tools for a simple task: stamping PDFs securely. Key features: • Detects whitespace automatically so stamps don’t cover text/images • Works mostly offline for privacy and security • Supports bulk stamping for multi-page PDFs • Lightweight, fast, and easy to use Would love to hear your feedback, what features...
StampPDFSecure, PDF stamping with smart auto-positioning
StampPDF lets you stamp PDFs securely without the cloud. It detects whitespace to avoid covering content, handles multi-page/bulk jobs, and keeps files small. Perfect for teams handling contracts, permits, or compliance docs.
StampPDFSecure, PDF stamping with smart auto-positioning