According to TechCrunch, OpenAI announced on Wednesday that it is launching an app directory for ChatGPT, swiftly dubbed an “app store,” and is now allowing developers to submit their apps for review. This follows the company’s initial announcement in October about bringing apps to its chatbot, which featured integrations from major platforms like Expedia, Spotify, Zillow, and Canva. The company states that these apps extend conversations by bringing in new context for actions like ordering groceries or creating slide decks. Developers can use the beta Apps SDK to build experiences and submit them through the OpenAI Developer platform to track approval. A number of approved apps are slated to start launching within Chat in the coming year. This is a significant step for OpenAI to expand ChatGPT’s ecosystem and give users more reasons to stay engaged.
The platform play
Here’s the thing: this was inevitable. Every major tech success story eventually tries to become a platform. Think iOS, Windows, Salesforce. OpenAI is basically doing the same thing with ChatGPT. They’ve got the users, now they need the “killer apps” that aren’t just made by them. Letting Expedia handle travel or Canva handle design makes ChatGPT infinitely more useful without OpenAI having to build all that complexity themselves. It’s smart. But it’s also a huge gamble.
The approval hurdle
Now, the big question is: what gets approved? OpenAI says developers can submit apps and track their status, but they’re being coy about the guidelines. This is where it gets tricky. Will they allow a rival AI model’s app? What about apps that are just thin wrappers for existing services? The approval process can make or break an app store. If it’s too restrictive, developers will walk. If it’s too loose, the user experience becomes a chaotic mess. And let’s be real, OpenAI doesn’t have a great track record here—remember the chaotic rollout of GPTs? This needs to be smoother.
Locking users in
So what’s the real goal? It’s all about retention and data. Every time you use the Kayak app inside ChatGPT to book a flight, you’re not going to Google. You’re staying in OpenAI’s walled garden. They learn more about your intent, your preferences, your behavior. That data is pure gold for refining their models and, eventually, for targeting ads or subscriptions. This move isn’t just about features; it’s about building a moat. The risk for users is getting locked into a single interface for everything, which sounds convenient until you think about choice and competition.
Can developers win?
And what’s in it for the devs? Right now, the monetization path is unclear. OpenAI hasn’t announced a revenue share model. Are developers just building free features to make ChatGPT more attractive, hoping for exposure? That’s a tough sell for anything beyond simple integrations. Big brands like Spotify can afford it for the distribution. But for a small startup, building for this platform is a bet that OpenAI will eventually cut them in on the profits. I’m skeptical. History shows platform owners usually take a hefty cut once they have the leverage. Developers should tread carefully and not put all their eggs in this one basket.
