Opik

Enables AI agents to browse and filter experiment runs, retrieve metrics, logs, tables, and artifacts, plot comparisons as PNG images, and download full-fidelity training history for local analysis against a self-hosted Open Train server. It also provides diagnostic tooling that surfaces metric evidence and investigation leads without asserting causality.

Provides read-only tools to generate test plans for AI agents, score agent runs, and produce reusable evaluation scorecards.

Scores AI agent trajectories and detects silent failures, loops, and reliability issues with zero external API cost.

Connects AI assistants to OpenTelemetry trace backends such as Jaeger, Tempo, Traceloop, and Datadog, enabling natural-language searching of traces and spans, error discovery, and service listing. It also provides LLM-specific analytics, including token usage aggregation, model performance comparison, and identification of the most expensive or slowest traces.