Open Source · MIT

The AI-Native
End-to-End Testing Platform

Write test steps in plain English. Iris runs them in a real browser, streams every frame live, and plugs straight into your CI — no selectors, no flaky scripts.

Features

Everything you need to ship with confidence

From writing your first test to gating production deploys — Iris covers the whole testing loop.

Tests in plain English

Describe each step the way you'd explain it to a teammate. The AI agent finds the elements and performs the actions — no selectors to write or maintain.

Live browser authoring

Build tests against a real browser you can see and control. Every frame streams live, with network, console, and storage panels right beside your steps.

AI test generation

Point Iris at a page and it explores the UI, identifies the key journeys, and drafts ready-to-run scenarios to review and save.

Flows & prerequisites

Chain tests visually on a canvas and reuse setup steps like login — state carries over in one browser session.

CI/CD ready

Trigger runs from any pipeline with the CLI or REST API. Exit codes reflect results, so failing tests fail the build.

Organized test suites

Nested folders, drag-and-drop, and bulk actions keep hundreds of tests manageable. Export suites as JSON and import them anywhere.

Variables & secrets

Store credentials once per workspace and reference them as {{variables}} in any step — secrets stay masked everywhere.

Rich run reports

Per-step screenshots, frame-by-frame recordings, and browser logs on every run. Edit the report and export it as Markdown.

Self-host in minutes

MIT-licensed and built for your infrastructure: one docker-compose for Postgres, Redis, and MinIO, plus local or cloud browsers.