Agenta-AI/agenta
An open-source platform for managing prompts, evaluating LLMs, and observing LLM application behavior in production.

Agenta provides an integrated LLMOps workflow combining a prompt playground, prompt versioning and management, LLM evaluation using various methods including LLM-as-a-judge, and observability tooling. It supports evaluating RAG systems and monitoring application reliability. The platform targets developers building LLM-powered applications and provides the infrastructure to iterate on prompts, run A/B tests, and track performance metrics over time.