Follow ZDNET: Add us as a preferred source on Google.

ZDNET’s key takeaways

• OpenAI is reportedly building a desktop superapp to unify its AI tools.
• Fragmented apps are limiting how often users like me use these tools.
• A unified app could reduce friction and improve agentic capabilities.

It sounds like OpenAI may be building something that could make it easier to find and access most of its AI tools in one place.

According to The Wall Street Journal, the company is planning a desktop “superapp” that combines ChatGPT, its Codex coding tool, and its Atlas browser into a single experience. The goal? To consolidate its line of AI tools and double down on agentic capabilities.

The report cites internal discussions and leadership changes, but OpenAI has not publicly confirmed any details or a launch timeline. CEO Sam Altman has also yet to directly comment on the so-called superapp on X, where he is typically very active, as of March 20.

Also: I tested GPT-5.4 Thinking, and it gave great answers (until I dove deeper)

Still, if true, it could be a step in the right direction, for me at least. Right now, using OpenAI’s products means juggling multiple apps. It’s honestly the biggest reason I don’t use some of them as much as I’d like, especially Atlas.

(Disclosure: Ziff Davis, ZDNET’s parent company, filed an April 2025 lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)

The real issue is friction

I test AI a lot for work, from Claude to Gemini, but ChatGPT is the one I keep going back to most. I use it to research, find deals, edit images with Adobe, place Instacart orders, make presentations, check for scams, and even have fun with my daughter by bringing her artwork “to life.” Do I think other tools do certain things better? Yes. But for most things AI, it’s where I start.

Also: I stopped using ChatGPT for everything: These AI models beat it

Unfortunately, over the last couple of years, ChatGPT has added some friction to the experience. I can no longer go to one website in my browser, or even one app on my desktop or phone, to access everything OpenAI offers. For instance, if I want help with automation, I need to switch to Codex. It’s incredibly powerful, but it’s an entirely separate app that I have to download.

Atlas is another example. When it first launched, I frequently tested it for shopping and different agentic capabilities, and it was a lot of fun. In one case, I even let it compile and place a Walmart order for me, and find the best tickets to Disney on Ice.

But I have stopped using it over the last few months for one simple reason: I live in Chrome on my desktop. Opening a separate AI browser requires me to break my normal workflow. Sorry, but I’m a creature of habit. Plus, Chrome is increasingly adding more Gemini capabilities. Recently, I used its Autopilot agentic feature to shop, research, and email for me.

Then there is Sora. To be clear, it’s not currently rumored to be included in OpenAI’s upcoming “superapp,” but that would be nice. I used it a lot at first, like when I animated my daughter’s artwork, and it’s pretty cool. But again, it lives on its own and is limited to my phone. Now I cannot remember the last time I opened it. Why must everything be separated?

I’m a creature of habit

The reason I don’t use Codex, Atlas, or Sora often is not because they lack features or capabilities. It’s because they require me to switch between apps. That might sound minor, but it adds up quickly. Every extra app is another download, another login, and another habit to maintain. I’m already trying to cut down on my screen time, and that includes limiting the apps I use.

So most of the time, I default back to what’s easiest. That means staying in one place, whether that’s ChatGPT in my desktop browser or the app on my phone. It also means I end up using only a fraction of what OpenAI has built. That’s the problem, and likely why OpenAI believes a superapp makes a lot of sense. Putting ChatGPT, Codex, and Atlas into one interface could remove the friction keeping many people, myself included, from using all of its tools. It would also align with where AI is going.

Also: AI agents of chaos? How bots talking to bots can go sideways fast

Agentic AI is designed to complete tasks across multiple steps, whether that’s researching, coding, or browsing. But wouldn’t AI agents work best if all those capabilities lived in the same place? One app to ask a question, write code, browse the web, generate media, and complete tasks without jumping around. The thing is, I want this everywhere, not just on desktop.

Bring the ChatGPT superapp to phones, too, OpenAI.

When will this actually happen?

I’m guessing it will happen in stages, not all at once.

The Wall Street Journal reported that top executives, including Altman, have spent the last few weeks reviewing OpenAI’s product portfolio and identifying areas to deprioritize. In a recent all-hands meeting, Chief of Applications Fidji Simo reportedly told employees they could not afford to be distracted by “side quests.” The company is also closely watching Anthropic, which has been gaining traction with enterprise and coding customers, and is now operating with a sense of urgency described internally as “code red.”

Also: How to switch from ChatGPT to Claude: Transferring is easy 

But isn’t building a superapp still a side quest, even if it is meant to consolidate all the others?

Nevertheless, based on the report, OpenAI is expected to roll out more advanced agent capabilities in Codex in the coming months before merging everything into a single experience. It’s focused on the desktop, with ChatGPT’s mobile app remaining unchanged.