An agent-powered QA tool that knows what to test without you telling it. Your very own junior QA that builds your tests from scratch and runs them to check your app for bugs. You build your app, we make sure it works.
@nishkarsh_srivastava Thank you so much. This is exactly the idea. Help people to get rid of the annoying part of e2e testing so that they can focus on building!
that is insane 🤯
the test cases it generated were pretty good! I needed to update a selector to be something less transient (it picked "has text: 178 images" in a image gallery 😂) but that was easy enough to do.
This could definitely make testing life less painful. Thanks!
@jonnymiles thanks for trying it out and providing feedback! Yes, we have to work on better selectors. The question for us is, what is the impact of a bad selector if you have auto-maintenance available? And can we make auto-maintenance understand that it fixed a selector for the 3rd time and that it should maybe choose a better one? Just some thoughts. We are on it :)
Let us know if you need any help with your testing!
@harry_sin1 so far we want to nail web apps at a level we're satisfied with and ship AI test auto-maintenance. Once we have that in place, mobile is the next step.
Report
Octomind is incredible. They've have finally built a QA agent that actually works. Marc and team are great to work with!
Congrats on launching @marc_mengler@germandrummer92@ma_zah@daniel_rodler and the entire Octomind Team! 🎉
An AI-powered testing tool for web apps that catches bugs before users do is such a valuable addition for developers. The fact that Octomind automatically writes and updates tests without needing manual input is a massive time-saver for teams, and the ability to integrate it into CI/CD pipelines makes it even more powerful. Exciting stuff!
A couple of questions about the future of Octomind:
Handling complex app testing: As your AI evolves, how do you plan to address the needs of more complex web applications with dynamic elements and integrations? Are there plans to train your AI to handle intricate, real-time data or user interactions, which tend to be more difficult to test?
Test coverage and customization: Does Octomind allow developers to customize or tweak the tests it generates, or is it purely autonomous? It would be interesting to know how much flexibility dev teams have to adjust the AI-generated tests to fit specific use cases or edge scenarios that are unique to their app.
For some feedback:
The concept of automating tests with AI is fantastic, but I believe offering clear test coverage reports (e.g., what’s been tested, what potential gaps exist, etc.) would be incredibly helpful. This would give developers greater visibility into how thorough the AI’s testing is and ensure that they’re covering critical areas of their apps.
Suggestion: Consider enhancing the platform with detailed test coverage analytics to allow users to review and feel confident that key aspects of their web apps are being thoroughly tested. This could help teams prioritize which areas need manual intervention or further scrutiny.
I also noticed there’s no live chat support on the site. Adding a customer support chatbot could provide quick answers for users setting up their tests or trying to integrate the tool into their CI/CD pipeline. It would improve onboarding and make troubleshooting much faster for new users.
Overall, Octomind is tackling a significant pain point for developers, and I’m excited to see how it grows from here. Keep up the amazing work! 🙌
@marc_mengler@germandrummer92@ma_zah@uvkewate
Complex app testing
As long as interactions stay inside the browser we will be able to handle them. There are of course areas which are more challenging like canvas or real time communication. We have to see how frequently people are asking for it. If a serious need arises, we are all in to tackle these challenges.
Customisation
We always start with a prompt. The prompt can either come from our system or you can enter your own prompt - or copy it from your test management system. This is the first level of customisation.
Once the first flow is generated you have full control over it. You can review and modify every single step it took. You can also add additional steps if the need arises. Or you can further refine your prompt and regenerate either all steps or you keep those which are already good. This is the second level of customisation.
Should you be unhappy with the selectors the system came up with you can easily directly enter Playwright locators. This is the third and most technical level of customisation.
Basically you are always in full control.
Coverage
I love the idea. Not an easy one, I have to say. But definitely super helpful if it can get solved in a good way. It is on our list.
Support Bot
Also a good idea. Thank you. I think we have to build up a more comprehensive knowledge base first to make it really helpful. I am sure we will install one once we are at this point.
Other than that: Thank you so much for sharing thoughts and feedback! This is much appreciated!
@arjunrajkumar I am glad you like it! If your mobile app is build on react native and accessible via URL, yes, it should work! But if it's a native Android or iOS app, we don't support it yet. On which of these are you developing your mobile apps?
Replies
Cortex
Octomind
PicLooks Avatars
Octomind
Octomind
Octomind
Humans in the Loop
Octomind
Octomind
Octomind
Octomind
Octomind
Humans in the Loop
Octomind
ElephantPath
Octomind