Launching today

Explorbot - autonomous web testing
AI checks your web app so QA can focus on complex scenarios
1 follower
AI checks your web app so QA can focus on complex scenarios
1 follower
Explorbot is an AI exploratory testing agent for QA leads working with limited team capacity. It explores your running web app, plans and executes checks, and records steps, logs and video for review. Use it for initial checks on new features and repeated checks as they change. It helps reduce time spent clicking through the app, so testers can focus on complex scenarios and product risks. Your team reviews the findings and makes the final call.




Free
Launch Team / Built With

AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted

Hey everyone!
I'd love to share the project I've been working on for about a year: Explorbot, an autonomous testing agent. For a decade I've worked on testing and created the test frameworks Codeception (for PHP) and CodeceptJS. But I didn't want to write test code myself. I wanted the code to test the product on its own. With AI, I could finally achieve this!
The work started a year ago. Unlike other approaches, I focused on a few key things: replayability, predictability, and separating planning from testing. Its architecture is a set of different agents combined into one predictable pipeline that explores a live website, analyzes it, writes a test plan, tries to execute the tests one by one, analyzes the results, and publishes them with videos.
We already run Explorbot on our products daily and on every PR.
Did I mention that Explorbot is open source with no vendor lock-in? I should have! Unlike all competitors, Explorbot is truly open source, and you can run it on GitHub Actions, GitLab CI, or Bitbucket Pipelines. So plugging it in can be quite simple!
Why did I decide to build an open source agent? Well, because I've built open source frameworks before, and what I learned is that the best solutions are free and open to contributions. I want as many people as possible to try Explorbot and give us feedback. Sure, websites are more complex than they look! And agents aren't always effective and can get lost in complex UIs. That's why improving Explorbot is our goal.
Explorbot is also really cost-effective, as it works with cost-efficient models like Haiku, GPT Luna, Muse Ember, and Flash models. It won't burn through your Ultra subscription on simple browser checks!
Eager to show it on Product Hunt, and curious which other product makers find it useful!