You describe
Write a test case as plain-English steps and an Expected Outcome, or let Self-driving write them. For example: search the Toolshop for hammer, and expect every result to be a hammer.
Documentation
Learn to test your web app with Testmode. Write AI browser tests in plain English, run them on any environment, let Self-driving explore your app and keep the suite up to date, and run tests from CI and AI agents.
Start here
Section titled “Start here”New to Testmode? The quickstart takes you from sign-up to your first verdict on a public practice shop, so you don’t need your own app yet. The Free plan needs no card.
Quickstart: run your first Testmode testOne guided path from sign-up to a first result.
What you can do with TestmodePlain-English tests, Self-driving, findings, schedules, the assistant, CI and MCP.
What Testmode can and cannot testA candid fit check: reachable URLs, logins, CAPTCHA and more.
A tour of the Testmode appThe sidebar, the menus, Overview and shortcuts.
How it works
Section titled “How it works”Testmode runs
A fresh cloud browser opens your environment, signs in when the test has a login, and carries out your steps as browser operations it works out once and replays every run.
You read the verdict
Each test ends Passed or Failed, with the reason, a recording of the browser, every operation it ran and the browser’s logs. Self-driving investigates what failed.
Testmode is an AI testing tool for web apps: no test scripts, no selectors and no changes to your code. Tests run in cloud browsers, so the app you test must be reachable from the public internet.
Browse by area
Section titled “Browse by area”ConceptsHow a test runs, how Self-driving works, and how organizations and projects fit together.
Account and teamSign in, invite people and understand the four roles.
Project setupProjects, environments, logins, variables and safety rules.
Test casesWrite steps and an Expected Outcome that run reliably.
Test plans and schedulesGroup test cases, run them on a schedule and email the results.
Runs and resultsStart runs, read verdicts, replays and logs, and track findings.
Self-drivingLet Testmode explore your app, keep the suite up to date and investigate failures.
The assistantAsk for tests, plans and schedules in a chat.
CI and AI agentsAPI keys, GitHub Actions, the REST API and MCP.
Plans and billingFree, the Pro trial, and how credits are counted.
Popular tasks
Section titled “Popular tasks”Turn on Self-drivingWhere it works, how much, and what it may do.
Work through the InboxAnswer questions, apply fixes, confirm bugs.
Write effective stepsWording that becomes the right browser actions, with before-and-after examples.
Test case examplesComplete test cases for search, cart, login, forms and checkout.
Log in automaticallyStore a test login once and reuse it across tests.
Run tests from GitHub ActionsFail the build when a test fails.
Connect an AI agent over MCPLet Claude, ChatGPT, Cursor or VS Code run your tests.
Troubleshoot failed testsWork out why a test failed and what to change.
Reference and help
Section titled “Reference and help”Run, test and exploration statusesWhat every status means, and why a test waits.
Limits and defaultsTime limits, waits, budgets, plan limits and retention.
GlossaryThe terms used across Testmode.
FAQAnswers to common questions.