opik
Comet tool for logging and evaluating LLM traces
Decision gist · record as of 2026-08-14
Yes. Opik is actively maintained with no known vulnerabilities and offers low install friction. It is genuinely open-source (Apache 2.0) with a self-hostable full platform, not just a client SDK. If you are building LLM agents or need observability and evaluation for LLM applications, this is a solid choice—especially if you want to avoid vendor lock-in or need to keep traces within your own infrastructure.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires Python 3.10 or later.
- The `opik configure` step sets up connection to a backend (cloud or self-hosted); without it, tracing may not persist.
- Low install friction with a pure-Python wheel.
License · maintenance · safety
Apache 2.0 License (permissive) — Apache 2.0 licensed and permissive. The full platform—server backend, web application, tracing, evaluation, and prompt management—is free to self-host with no data leaving your environment, not just the client SDK.
last release 2026-08-14 (0 days) · last repo commit 2026-08-14 · 21,388 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 3,695,539 downloads/mo, #2,523 on PyPI
Alternatives
Verify before relying
pip install opik
opik configure
from opik import track
@track
def my_function(input: str) -> str:
return input
my_function("test")- Whether the self-hosted platform deployment requires additional system dependencies or infrastructure beyond the Python package.
- Performance characteristics and scalability limits when self-hosting at scale (the description mentions 40M+ traces/day capability).
- Whether all integrations mentioned (Google ADK, Autogen, Flowise AI) are included in this package version or require separate setup.
What it is and what it does
Opik is a Python SDK and full-stack platform for observing, evaluating, and optimizing LLM applications and AI agents. It wraps functions with a `@track` decorator to automatically log traces of LLM calls, nested function calls, and agent activity, capturing full execution trees for multi-step workflows. The platform includes evaluation tools—datasets, experiments, and LLM-as-a-judge metrics—for testing hallucinations, moderation, and RAG quality. It also provides prompt management, a playground for experimentation, CI/CD integration via pytest, and production dashboards for monitoring feedback scores and token usage.
The package is built by Comet and released under Apache 2.0, meaning you can self-host the entire backend in your own infrastructure. It supports modern Python versions (3.10–3.14) and integrates with popular LLM frameworks. The SDK depends on litellm, openai, pydantic, httpx, and other standard libraries for tracing, configuration, and API communication.
Use it for
- Log and inspect full execution traces of multi-step agents and LLM pipelines during development and in production.
- Evaluate LLM outputs for hallucinations, factual accuracy, and moderation using LLM-as-a-judge metrics.
- Test RAG systems by measuring answer relevance and context precision across datasets and experiments.
- Integrate LLM evaluation into CI/CD pipelines using pytest to validate prompts and models on every commit.
- Monitor production LLM applications by tracking trace counts, token usage, and feedback scores over time.
- Manage and version prompts, then experiment with different models and parameters in a shared playground.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes.
Opik is actively maintained with no known vulnerabilities and offers low install friction. It is genuinely open-source (Apache 2.0) with a self-hostable full platform, not just a client SDK. If you are building LLM agents or need observability and evaluation for LLM applications, this is a solid choice—especially if you want to avoid vendor lock-in or need to keep traces within your own infrastructure.
Install
opik on PyPI
Before you install
Low install friction with a pure-Python wheel. Active maintenance with a recent release. Requires Python 3.10+. Depends on 19 runtime packages including litellm, openai, and pydantic, which are widely used and stable.
Requires Python 3.10 or later. The `opik configure` step sets up connection to a backend (cloud or self-hosted); without it, tracing may not persist.
License in practice
Apache 2.0 licensed and permissive. The full platform—server backend, web application, tracing, evaluation, and prompt management—is free to self-host with no data leaving your environment, not just the client SDK.
Quickstart
pip install opik
opik configure
from opik import track
@track
def my_function(input: str) -> str:
return input
my_function("test")Verify before relying
- Whether the self-hosted platform deployment requires additional system dependencies or infrastructure beyond the Python package.
- Performance characteristics and scalability limits when self-hosting at scale (the description mentions 40M+ traces/day capability).
- Whether all integrations mentioned (Google ADK, Autogen, Flowise AI) are included in this package version or require separate setup.
Package facts
| License | Apache 2.0 License permissive |
| Python support | Supports the current Python release >=3.10 |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 19 packagesboto3-stubsclickhttpxrapidfuzzlitellmopenaipydantic-settingspydanticpytestrichsentry_sdktenacitytqdmuuid6jinja2watchfilestree-sittertree-sitter-javascripttree-sitter-typescript |
| Maintenance | Actively maintained 0 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 3,695,539 / month, #2,523 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 2 - Pre-AlphaIntended Audience :: DevelopersLicense :: OSI Approved :: Apache Software LicenseNatural Language :: EnglishProgramming Language :: Python :: 3Programming Language :: Python :: 3 :: OnlyProgramming Language :: Python :: 3.10Programming Language :: Python :: 3.11Programming Language :: Python :: 3.12Programming Language :: Python :: 3.13Programming Language :: Python :: 3.14 |
Evidence: opik-2.2.29-py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “llm-as-a-judge evaluation”
- opikOpik is an open-source LLM observability and evaluation platform that…
- langfuseLangfuse is a Python SDK for observability, tracing, and evaluation…
- strands-agents-evalsStrands Evals SDK provides a comprehensive evaluation framework for…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also arize-phoenix · langfuse · mlflow-tracing · trulens · traceloop-sdk · mlflow · judgeval · promptflow-devkit · regopy · evidently