Google pulls AI Test Kitchen from app stores

Google is cleaning house again, and its experimental AI playground just got kicked off your phone.

Google has quietly pulled its AI Test Kitchen app from both the Google Play Store and Apple App Store. The company later confirmed the move to 9to5Google, explaining that it is ditching native mobile apps to focus entirely on the web.

The kitchen is down to one burner

Google launched AI Test Kitchen last year as a public sandbox for curious users to test frontier models like LaMDA 2. Early demos were quirky and controlled, letting people break complex goals into sub-tasks or talk endlessly about dogs to test if the model could stay on topic.

Google says sticking strictly to the web makes shipping updates far easier than wrestling with two separate app stores. But the writing has been on the wall for a while.

Last November, Google promised a full Season 2 of experiments that never actually materialized. Today, the Test Kitchen hosts just a single, lonely demo: MusicLM, the text-to-music model announced at Google I/O.

Classic Google product confusion

To make things even more disjointed, Google unveiled another experimental front door called Google Labs back in May. That hub is where users now sign up for generative AI tests across Search, Workspace, and the NotebookLM note-taking tool.

The weird part is that MusicLM is also listed inside Google Labs. Having two separate websites to showcase overlapping AI experiments is peak Google, and shutting down the mobile apps is likely a way to stop burning engineering cycles on redundant storefronts while competing with OpenAI and Anthropic.

Why it matters

For everyday tinkerers, you aren't really missing anything. If you want to mess around with MusicLM, you can still fire up a desktop browser and do exactly what the mobile app let you do. There is no major feature loss here.

For the broader AI landscape, this is a clear sign that Google is moving past the stage of polite public exhibits. When the AI Test Kitchen debuted, Google was treating generative AI like a fragile museum piece that needed heavy guardrails. Now that ChatGPT and Claude have turned AI into an existential battleground, boutique demo apps are a distraction from revamping Google's flagship search and enterprise tools.

We think dropping mobile builds is the right technical call, since AI models iterate too fast for app store review cycles. However, the move also highlights Google's persistent organizational mess. Running two half-baked experimental portals side by side looks less like a deliberate strategy and more like internal teams failing to coordinate their sandboxes.