LLM Observability

Open source LLM observability platforms that trace, monitor, and evaluate AI applications — capturing prompts, model calls, retrieval steps, and agent actions so you can debug behavior, score quality, and track cost and latency instead of flying blind in production.