Evidently

Evidently is an open-source framework for evaluating, testing and monitoring machine learning and LLM-powered systems. It computes more than 100 metrics covering data quality, data drift, model performance and GenAI output quality, and turns them into readable reports and test suites. The project has about 8,000 stars on GitHub, is released under the Apache 2.0 license, and was actively updated in 2026.

Overview

Evidently works across the whole lifecycle of an AI system, from tabular data pipelines to generative AI applications. You can run one-off evaluations in a notebook, add pass/fail test suites to a CI workflow, or feed results into a monitoring dashboard that watches a production model over time. Results render as interactive HTML reports that are easy to share with teammates, and the same checks can run on a schedule against live data. For LLM applications, Evidently adds text-focused evaluations and descriptors on top of the classic ML metrics, so RAG systems and chatbots can be tested with the same tooling as a classification model.

Key Features

• 100+ built-in metrics for data drift, data quality, model performance and LLM outputs
• Interactive HTML reports and visual dashboards that work from Jupyter notebooks
• Declarative test suites with conditions, suitable for CI/CD quality gates
• Monitoring service for tracking evaluation results over time in production
• Support for tabular, text and embedding data, from classic ML to GenAI pipelines
• Integrations with tools such as MLflow and workflow orchestrators
• LLM evaluation helpers including text descriptors and prompt-based scoring

💳 Pricing

Freemium. The Evidently library is free and open source under Apache 2.0. The team also offers Evidently Cloud, a paid hosted platform for collaborative monitoring and managed evaluation workflows.

Who Is It For

ML engineers, data scientists and LLM application teams who need systematic evaluation and production monitoring for their models without building the observability stack themselves.