🎥 Free on-demand webinar "How to use LLM-as-a-judge to evaluate LLM systems." Get your access

Open-source evaluation and observability for ML and LLM systems

Evaluate, test, and monitor AI-powered systems. From tabular data to LLMs. Built for data scientists, AI, and ML engineers.

Explore docs

GitHub

Open-source

Evidently Python library

Evidently offers 100+ pre-made metrics and a framework for custom ML, LLM, and data evaluations.

You can start with ad hoc reports in a minute, run automated pipeline and CI/CD checks, or deploy a complete monitoring stack with a live dashboard.

5000+

GitHub stars

20m+

Downloads

2500+

Community members

Join 2500+ ML and AI engineers

Get support, contribute, and chat about AI products.

Join our Discord community

Reports

Preset evaluations

Run instant pre-configured ML and LLM evaluations. Get visual reports or export the results elsewhere. Swap and add metrics easily.

Test suites

Configurable Test Suites

Run structured checks and get alerted on failures. Debug test results with visual reports.

Zero setup option. Auto-generate test conditions based on past examples.

Trigger actions. Use test outcomes to set up alerts, initiate retraining, or stop pipelines.

Get started

Dashboard

ML monitoring dashboard

Track the health of all ML models and datasets from a central interface.

Batch or real-time integration. Send data from live services or in batches.

Customizable panels. Visualize statistical data summaries, test results, or specific metrics.

Get started

Cloud vs. oss

Evidently Cloud for teams and enterprises

Built with an open-source foundation, Evidently Cloud is the easiest way to implement AI observability. Get a collaborative workspace for your entire AI product team.

Start free

Scalable backend: we handle it.