OpenAI hasn't turned ChatGPT into a new App Store yet, and developers are complaining

OpenAI wanted to turn ChatGPT into a new hub for digital services, a place where users don't just ask questions but also order rides, listen to music and book hotels without leaving the chat. Six months after the built-in mini-apps launched, though, the plan is stalling: the catalog is growing, but the convenience and value for users and developers remain limited.

Ambitions versus Apple

When OpenAI opened a platform for third-party services inside ChatGPT, the bet looked big. Companies such as Spotify and Booking could embed their mini-apps directly in the conversational interface. The idea was simple: if users already spend time in ChatGPT, it makes sense to let them take the action they need right away, whether finding a ticket, choosing a hotel or opening a music service, without switching between separate apps and tabs.

The design echoed the early logic of the App Store: first a platform with a broad audience appears, then a developer ecosystem grows around it, and eventually the entry point itself becomes more valuable than any single app. For OpenAI, it is also a way to reduce its dependence on other companies' platforms, above all Apple, which controls access to users on the iPhone and takes a commission on digital services.

Why the start is stalling

On paper, there is progress: in six months, ChatGPT has gained more than 300 integrations. But having a catalog has not turned it into a place people use every day. According to developers, the mini-apps are not particularly well placed, and their capabilities are often trimmed down. A user may see an attractive integration shell but not always get the full workflow that would justify opening the service through chat in the first place.

The partners' list of complaints is fairly down-to-earth:

• A long, tiresome app approval process
• Raw, in places buggy development tools
• A lack of data on usage and user behavior
• Limited functionality of the integrations themselves
• Poor visibility of apps inside ChatGPT

For developers, this is a painful combination. When a platform offers little analytics, it is hard to tell what people launch, at which step they drop off and which features are worth improving. And if the audience also struggles to find the apps, teams have a harder time proving internally that a ChatGPT integration is worth the time and budget.

Where the model broke down

The main problem seems to lie not only in OpenAI's tooling but in the very structure of such a market. Partners are in no hurry to give OpenAI the customer relationship, and especially payments. For any large service, this is no longer just a technical integration but a question of control over audience, revenue and brand. If a user completes a purchase or subscription inside someone else's interface, the owner of that interface eventually becomes the main intermediary. That is why companies connect cautiously and often keep their integrations on a short leash.

As a result, OpenAI wants to build an App Store-like ecosystem but without one key element: participants' willingness to build truly full-featured products on the platform. It becomes a vicious circle. Users see little benefit and rarely open the mini-apps. Developers see no traffic and are in no hurry to invest more deeply. The platform is growing in partner count but not in the strength of habit.

What this means

The mini-apps story shows that turning a popular chatbot into a full platform is harder than simply opening a storefront of integrations. It takes not only partners and a catalog but also clear distribution, stable tools, transparent analytics and businesses' confidence that they won't lose the customer inside someone else's shell.

For now, OpenAI's project looks more like a strategic claim on the future than a real competitor to the App Store in its current form.