SpecGuard is a command-line tool that turns AI safety policies and behavioral guidelines into executable tests. Think of it as unit testing for your AI's output. Instead of trusting that your AI will follow the rules defined in a document, SpecGuard enforces them.

QA framework for testing LLM agents with cryptocurrency data support. Includes Mock LLM for offline testing, FastAPI endpoints, Bitcoin data validation, and PyTest test suite. Built with Python, LangChain, and FastAPI.

A production-style FastAPI service that enforces structured, safe, and deterministic LLM outputs using versioned prompts, strict schema validation, and automated retry with repair prompts.

🛡️ Enforce AI behavior guidelines with SpecGuard, a tool that turns policies into executable tests for reliable and scalable AI output management.