Comet / Opik
comet.comThe Fastest Path to Agents That Work
AI Toolsllm-observabilityagent-evaluationmlopsexperiment-trackingllm-tracingopen-sourcegenai
About
Comet is an end-to-end AI model evaluation and observability platform, featuring Opik — an open-source LLM tracing and evaluation tool for GenAI apps and agents. It enables developers to log traces, run evaluations, annotate results, and automatically fix agent code via a built-in coding agent called Ollie. The platform also includes MLOps capabilities for experiment tracking, model versioning, and production monitoring.
Problem
Developers building GenAI applications lack visibility into how their LLMs and agents behave, making it hard to debug, evaluate, and improve them reliably.
For
ML engineers, data scientists, and developers building LLM-powered applications and AI agents
How it works
Opik instruments LLM calls and agent steps via a few lines of code, logging traces and eval results that feed into automated scoring, human annotation, and an AI coding agent that writes fixes directly to the codebase.
Business model
freemium
Status
launched
Company
Comet