guardrails-ai

Deterministic cost / loop / time budgets · full observability · crash-resumable runs · human-approval gates · a memory you own. Self-hosted. Your keys. No telemetry. Point it at your existing agents - one env var.

LLM Testing Library for Java: Fault Injections, Prompt Injections, Adversarial Attacks, Guardrails Testing, Bias Testing

📚 Stateful, self-correcting RAG research assistant built with LangGraph & Streamlit. Features hybrid search (dense + BM25), Cohere reranking, Tavily web/arXiv claim verification, session-isolated Qdrant storage, Guardrails AI, DeepEval testing, LangSmith tracing, /btw off-topic channel and AWS Fargate deployment configured with Terraform & CI/CD.

Python port of QFIRE: a declarative, positive-security prompt firewall. Rules and chains are authored as YAML; the engine evaluates them cheapest-first with short-circuiting, fails closed on any detector error, and writes an auditable decision trace for every evaluation.

Make your AI write shippable Verilog — a ruleset + a checker for Xilinx FPGA RTL

Content Guardian — Intelligent AI Middleware for content moderation and validation. Powered by Guardrails AI to ensure LLM safety and structured output via a modular API.

Automated red-teaming and safety layering for a RAG agent — Promptfoo attack generation, Guardrails AI validation, and Llama Guard classification, with before/after break-rate comparison.

A lightweight, external file integrity monitor (FIM) to protect your project architecture and guardrails from silent overwrites by autonomous coding agents.

A full-stack AI financial chatbot built on NVIDIA's API with dual safety guardrails, RAG over research documents and live market tools.