We are using cookies for secure log-in and to improve our app.
First party only, no data krakens 🦑 Learn more in our privacy policy.
You’ve sent us heaps of feedback on prompt-based test generation. Creating tests in plain English is great, but writing the right prompt proved tricky - our best practice docs only got you part of the way.
We are now stepping in. The Agent now refines your prompt automatically, adding the detail and precision needed for rock-solid tests. And if the “autotuned” prompt drifts, you can jump in, edit it, and steer the Agent back on track.
Introducing better prompts:
You never signed up to be a full-time prompt engineer - and now you don’t have to. We’ve tuned the Agent to clean up and enrich whatever natural-language prompt you give it.
Take this example of an eshop checkout test. Obviously, “go to checkout” needed some work.
The Agent expanded that prompt into a structured prompt with three clear sections:
Less guessing, more green ticks.
Select a folder to store your test, add tags, and a change prompt details if needed. We want you to stay in control of the process.
Go to app and prompt away, you will notice the difference in the test creation experience.
More blogs! Kosta’s first article is live, unpacking why we finally let arbitrary waits into our tests. We’ve railed against developer dogma before, but dropping this rule took some soul searching.
---
We iterate on the better prompt feature. This is the perfect time to give us your take and wishes. I’m here to listen, every input is highly appreciated!
There's one more product update coming next week to wrap up the July feature bonanza.
