Opik
Free tierOpen-source AI observability & evaluation platform for the agentic era
Free tier available·Technical·API available·Open source
Key strengths
End-to-end agent trace logging and visualization30+ LLM-as-a-Judge evaluation metricsAutomated code fix suggestions via Ollie coding assistantTrue open-source with self-hosting supportReal-time production monitoring with guardrails and cost tracking
Free tier + paid plans
United States
Self-hostable
No ratings yet
Opik provides full-stack LLM observability through trace capture, structured evaluation pipelines, and production monitoring. It supports 30+ LLM-as-a-Judge metrics covering answer relevance, hallucination, context precision, and more, alongside a plain-text assertion system for defining unit tests without writing individual eval metrics. The platform includes Ollie, an AI coding assistant that reads traces, suggests fixes, and writes them directly to the development codebase with built-in version control and regression testing. Opik is fully open-source (hosted on GitHub under the comet-ml org) and can be self-hosted or used via the managed Comet cloud.
