OpenAI consolidates ChatGPT, Codex, and browser into one desktop super app

OpenAI is done pretending that having multiple separate products is a feature. The company is merging ChatGPT, its Codex coding platform, and its standalone browser into a single desktop application. Fidji Simo, who heads OpenAI's applications division, confirmed the direction publicly. The goal is one unified interface built around handling complex, multi-step tasks without the user needing to switch between tools.

This isn't a minor UI refresh. Consolidating three distinct products into one desktop app is a structural bet that users want fewer apps doing more work, not a growing collection of specialized tools. OpenAI has been watching its enterprise market share get chipped away by Anthropic, and the response is a platform play rather than a product-by-product competition.

Why a super app, and why now

The timing matters. Anthropic has been making steady progress with Claude in enterprise settings, particularly among teams that care about reliability and context handling in long workflows. OpenAI's answer isn't to out-feature Claude point by point. It's to change the frame entirely by offering a single desktop environment where an AI agent can browse the web, write and run code, and hold a conversation without any manual handoff between apps.

Agentic task handling is the core idea here. Instead of a user opening ChatGPT to ask a question, then switching to Codex to write a script, then opening a browser to research something, the super app is supposed to handle that entire chain in one place. The AI takes on longer tasks with less step-by-step instruction from the user. That's a meaningful shift in how OpenAI is positioning its product.

What happens to the mobile app

OpenAI has said the mobile app will stay separate. That makes sense practically. Desktop and mobile usage patterns are genuinely different. Agentic workflows that involve running code or extended multi-step tasks are much better suited to a desktop environment. Keeping mobile as its own product means OpenAI isn't forcing a compromised experience onto phone users while it builds out the desktop platform.

It also signals where OpenAI sees the highest-value use cases for agents. Power users, developers, and enterprise teams tend to work on desktop. That's the audience most likely to benefit from a combined coding, browsing, and conversation tool. Mobile remains the access point for lighter, faster interactions.

Product sprawl was becoming a real problem

OpenAI has launched a lot of products in a short time. ChatGPT, Codex, Sora, the browser, various API tools and plugins. For individual users and businesses trying to build on top of OpenAI's stack, the number of entry points was getting confusing. A super app reduces that friction significantly. One login, one interface, one place where the AI can actually do end-to-end work.

There's also a competitive logic here beyond Anthropic. Google has Gemini integrated across its entire product suite. Microsoft has Copilot embedded in Windows and Office. Both of those are platform-level integrations. OpenAI, despite being the company that started the current wave of consumer AI interest, doesn't own an operating system or a productivity suite. A desktop super app is the closest it can get to that kind of platform presence without buying or building something entirely new.

What this means for Codex users specifically

Codex was already more of a backend API product than a consumer-facing tool. Developers accessing it through the API probably won't notice much change in how they integrate it into their own systems. But for users who interacted with Codex through OpenAI's own interfaces, it will fold into the super app experience. Code generation, execution, and debugging would become part of the same environment where you're also having conversations and browsing.

That kind of integration is exactly what a lot of developers have been building manually with custom setups. Having it built into a single product from OpenAI removes a layer of setup work, assuming the execution environment is reliable enough for real development tasks.

The browser piece is the wild card

OpenAI's browser product has been the least visible of the three being merged. Building a browser and getting users to actually switch from Chrome or Safari is a historically difficult problem. But inside a super app, a browser doesn't need to replace your default browser for everything. It just needs to be good enough for the AI agent to use it on your behalf when completing a task. That's a much more achievable bar.

If the agent can open web pages, fill in forms, extract information, and return results without the user leaving the super app, then the browser component justifies itself. The user might never think of it as a browser at all. It becomes just another capability the AI has access to during a task.

Where things stand

OpenAI hasn't given a hard launch date for the fully merged desktop super app. Fidji Simo's confirmation establishes the direction without locking in a timeline. Given OpenAI's release cadence over the past year, a public beta could arrive within months rather than years. The company has been shipping at a pace that makes long development timelines unusual.

For now, the separate products continue to operate as they do. But the strategic direction is clear: OpenAI wants to be the single desktop environment where AI work happens, not one of several tools you use alongside others.