Wren / QA testing

Stop babysitting your end to end tests.

Describe your test in a sentence or two. Wren runs your tests in parallel. Seldon's behavioral models guide what each agent tries next, so one workflow can reveal test cases you never thought to write.

First 10 runs free $3 per completed run after that.

Built for Engineers and QA teams

The question: What did this release break?

What you examine: Recorded runs, step-by-step, with evidence for every failure.

Your tests shouldn't be a project within your project.

Both tests create the same featured release and check its URL and heading. The scripted example uses Playwright to specify each browser action. With Wren, you describe the task and the result you expect.

Scripted browser tests The Old Way

Wren The New Way

Complete Wren test definition

How it works

Find the tests you didn’t know to write.

In the run below, a shopper clicked “Add to bag,” saw no confirmation, and tried again. Two items ended up in the cart. That gives you a new case to test: after the first click, can the shopper tell the item was added?

1. Write

Describe in plain English.

Describe the task as you would to a teammate. Include where to start and what a successful result looks like.

2. Run

Run your tests in parallel.

Agents choose their next step based on what they see. If a button gives no feedback, an agent might retry or look elsewhere.

3. Find

See what your tests missed.

Follow the recorded steps to see what went wrong, how the agent got there and how to reproduce the failure.

Pricing

Your first 10 runs are free.

Try a workflow your team already tests. Then pay only for completed runs.

$3 per completed test run.

Everything included

Model usage, browser execution, findings and recorded evidence. No separate token charges.

No seat fees

Your first 10 completed runs are shared across your workspace.

Pay for completed runs

Runs that cannot finish because of an agent or platform failure are not charged and don't use your free allowance.

Try Wren alongside your existing suite.

Pick one flow your suite already tests. Run it with Wren and review the recording to see which other paths it explored.

Get started free