Framer AI AgentsDesign and publish professional sites with AI
Promoted
Drizz
Mobile tests that write, run, and fix themselves
1.4K followers
Mobile tests that write, run, and fix themselves
1.4K followers
Drizz is an AI-powered mobile test automation platform built around intent-based testing. Simply describe what you want to test in plain English, Drizz executes it on a real device using Vision AI and automatically authors a reusable test case. No scripting, no flaky selectors, no manual maintenance. It adapts to dynamic UIs, integrates with your CI/CD pipeline, and gives your team reliable end-to-end coverage without the overhead.
Upstream
How does one install it?
Drizz
@louislecat Hey, you can download it right here: https://www.drizz.dev/download-desktop-app
Drizz
@louislecat You can install using this link: https://www.drizz.dev/download-desktop-app
can we try it?
Drizz
Yes@sudo_tallan !
Thanks for checking in.!
You can download the app and start right away!: Download the app from here: https://www.drizz.dev/download-desktop-app
Happy testing!!
@zaid_ahmed_ansari thanks ! I hope my QA engineer had a good discovery call last week.
forgot to ask you... US and Polish Apple stores?
Drizz
@sudo_tallan any app can be tested. Drizz is more of an agent that works on the device and is app agnostic.
you can get the ipa files of the apps you want to test , or just connect your xcode apple simulator where the app is installed with the tool and you are good to go.
further read: https://docs.drizz.dev/getting-started/drizz-desktop-app/drizz-desktop-app-setup/prerequisites/system-requirements-for-ios
@zaid_ahmed_ansari Thanks! It sounds promising. I wanted my team to explore your tool. We could do hands on testing in 3+ weeks with our Frank AI mobile app.
Brila
Congrats with the launch! How can one with pipeline of Maestro + Claude Code benefit Drizz?
Drizz
@visualpharm Thanks Ivan!
Great question — here's how I think of the stack:
• Claude Code helps your devs ship features faster.
• Maestro lets you script flows in YAML — works great until the UI shifts and selectors break.
• Drizz sits on top with Vision AI. You describe tests in plain English, they run on real iOS + Android devices, and they self-heal when your UI changes — no selector babysitting.
So in a pipeline, you'd typically use Claude Code for code, keep Maestro for the simple flows you already have, and bring in Drizz for the dynamic/critical flows where maintenance kills you (login, payments, checkout, search). Many teams move flaky Maestro suites to Drizz first and keep the stable ones as-is.
Happy to jump on a quick call if you want to map it to your tech stack specifically.
Curious how it handles multi-step flows where the UI state depends on previous actions — like a checkout that changes based on what's in the cart. Does the intent description need to account for that, or does Vision AI figure it out contextually?
@hirogure That’s a thoughtful question, and honestly one of the core challenges in building reliable agents.
The way to think about it: the agent works the way a person would. You give it the goal, it looks at the screen, decides what makes sense next, does it, then looks again. It doesn't care whether the cart has 2 items or 12, whether a coupon banner showed up, whether a delivery slot picker appeared. It just sees the current screen and reacts to it, the same way you would.
Between turns, we keep a sense of what's happened so far, what was tried, what changed, and what's still pending, across a few layers of memory so the agent stays coherent on long flows instead of forgetting the thread. After each action it looks back at the screen to confirm the step actually did what it intended, and re-plans if not.
For testing, the strictness comes from how you write the intent. A loose goal like "checkout the cart" lets the agent pick any reasonable path. A scripted intent like "open cart → apply coupon SAVE10 → choose UPI → place order" gets followed as written. The agent still observes and verifies each screen, but it doesn't get to wander off the path. Same loop underneath, different amount of freedom. You decide which one fits the test.
@amannnn That memory layer across turns is the part I was most curious about — good to know it's not stateless. The "loose vs scripted intent" framing makes sense for test design. Does the memory persist across separate sessions, or does it reset each run?
Sipcode
Hey, spent some time on Drizz's page and the no-script mobile testing angle is what pulled me in. one thing I kept thinking about, how does the agent handle non-deterministic UI like loaders, ads, or A/B variants? that's usually where script-free testing falls apart in my experience.
Flexprice
Drizz
@manish_choudhary19 The "let me just buy one more phone" struggle is SO real 😅
You won't need to anymore, Drizz handles all of that for you!
Thanks so much, Manish 💜
"Testing was a tax, not a tool" is a great line, and the 12px button story is painfully relatable. As a fellow builder in the automation space, I love that you went vision-first. Brittle XPaths have wasted more of my life than I'd like to admit.
Congrats on the launch, Yash, rooting for the whole team 🚀
Drizz
@varunrai the vision-first bet gets more interesting from here. Right now the heavy reasoning runs where the compute is, but the direction we're betting on is smaller, sharper models that run on the device itself — testing that's faster because it isn't waiting on a round trip, and more accurate because the model is specialized for screens rather than general-purpose.
Means a lot coming from someone building in the same space — rooting for your work too.