Skip to content

Quickstart

OttoTester tests web apps with AI agents — they explore your app, write the tests, and run them. This walkthrough takes you from a fresh account to a first passing test. It’s about ten minutes, most of which is the agents working.

Work in OttoTester happens inside a workspace. If a teammate invited you, you already have one — open it. If not, create one: see Set up a workspace.

Inside the workspace, create an application — the product you want to test — and add a variant for the environment to test against, with its URL.

That’s the minimum OttoTester needs: something to point at.

If the part of your app you want to test sits behind a login, set up a fixture so the agents can sign in. You provide credentials once; OttoTester handles the rest. See Set up authentication.

If the pages you’re testing are public, skip this step.

Start a run from the variant. A run uses a Test Authoring Template for its configuration, and the default template is fine for a first run. You can add an Agent Prompt to steer what the agents focus on — “test the checkout flow” — or leave it blank and let them explore.

Launch it. Now the agents work:

  • the planner explores the app and decides what to test,
  • the generator writes a Playwright script for each test,
  • the executor runs them.

When the run finishes, open its report. You’ll see each test and whether it passed. Open a test to see its steps; open a failed one to see exactly where it broke. See Read a run report.

That’s a working test suite — generated, run, and reported, without writing a line of code.