Skip to content
← Tools / Skills
ToolComet/10 trustVerified
Opik logo

Opik

Open-source LLM observability, evaluation, and agent tracing.

Visit official site →

About

Opik is Comet's open-source platform for LLM observability and evaluation: trace LLM calls and agent runs, run LLM-as-a-judge evaluations, manage prompts and datasets, and monitor production — across the full application lifecycle. 22.5k stars, Apache-2.0, free to self-host the entire platform, 1.74M PyPI downloads/month. It sits between Langfuse (lighter) and LangSmith (SaaS) — choose Opik when you want open-source with a serious evaluation suite. Comet's commercial cloud is the monetization path, which shapes the roadmap.

What it can do

  • **Tracing**: LLM calls, chains, and agent runs with full payloads
  • **Evaluation**: LLM-as-a-judge metrics, datasets, experiments
  • **Prompt management**: versioned prompts with playground
  • **Production monitoring**: dashboards, alerts, feedback loops
  • **Integrations**: OpenAI, Anthropic, LangChain, LlamaIndex, and more
  • **Self-hostable**: Apache-2.0, full platform via Docker

Setup

```bash

pip install opik

```

```python

import opik

opik.configure(use_local=False) # or self-host the platform

from opik.integrations.openai import track_openai

client = track_openai(OpenAI())

```

Self-host via Docker Compose; docs: comet.com/docs/opik.

Prerequisites

Python 3.9+. Opik Cloud account or a self-hosted Opik server.

Use cases

observabilityevaluationtracingprompt management

Works with

customGeneric APIlangchain

Trust score — 9/10

/10 trustVerified

Scored 9/10 (verified tier, metrics-v1). Evidence: 22,533 GitHub stars (log-scaled vs catalog peers); blended with ~1,741,711 monthly package downloads; last push 0d ago; 93 releases/12mo; 216 open issues. Manual review: Comet's OSS LLM observability/eval (tracing, LLM-judge, prompts, monitoring); 22.5k stars, 1.74M PyPI/mo; full self-host heavier than Langfuse; Comet shapes roadmap. License: Apache-2.0.

★ 22,533 GitHub stars

How the trust score is built

Full methodology →
Adoption · stars/downloads, log-scaled vs peers
8.4/10
Maintenance · recency, releases, issue load
9.6/10
Openness · license + docs
10.0/10
Manual review · sanity read + red-flag check
8.0/10
  • · 22,533 GitHub stars (log-scaled vs catalog peers)
  • · blended with ~1,741,711 monthly package downloads
  • · last push 0d ago; 93 releases/12mo; 216 open issues
  • · manual review: Comet's OSS LLM observability/eval (tracing, LLM-judge, prompts, monitoring); 22.5k stars, 1.74M PyPI/mo; full self-host heavier than Langfuse; Comet shapes roadmap

Sources

Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.