ChatGPT Atlas

The browser with ChatGPT built in

Mac

Artificial Intelligence

Love it! I used to keep a tab pinned with ChatGPT always open anyways and this makes it so much easier, plus the load time seems to be faster with the built in browser vs using on Chrome. And the ability to read and interact with any open site directly on the screen is great

I was using Comet from @Perplexity for a while, but I ultimately just went back to Chrome. It never really handled tasks fast enough or to a level of quality that I wanted. Maybe I'll end up giving this one a go though.

I came into Atlas seriously hyped after hands‑on time with @Dia Browser and @Perplexity 's Comet. And yes—Atlas clearly “gets” multi‑step tasks. You can feel the maturity of an agent that’s spent time manipulating real browsers from inside containers.

But it also feels like a browser that doesn’t understand the basics of being a browser:

- Extensions: You can install them, but they’re invisible in the UI. That makes simple stuff—like logging in with a password manager—basically impossible. In 2025. Awkward.

- Import: No Chrome, no Comet, no Dia. It’s Safari or nothing. That’s a head‑scratch when you’re courting power users with established stacks.

It reads like a move to reduce server bills by offloading more work into a local browser… while overlooking the core ergonomics that make a browser your daily home.

Net: impressive agentic brains, missing everyday browser bones. If Atlas is going to be the place I live on the web, it needs to respect the muscle memory and essentials of modern browsing—starting with visible extensions and sane imports.

The friction of switching browsers is too high. Hope that OpenAI has made it simple for us to just login and sync our data.

It’s a bit more limited than I initially expected, but I’m really curious to see how far it will go from here. I currently work with four monitors in my office , one of them is always running ChatGPT. If ChatGPT Atlas continues to evolve, I could honestly see myself getting down to just three. 😄 The potential here feels huge, and I’m excited to see how it grows.

I've noticed room for improvement though if I may share. I believe to have a frictionless experience, OpenAI needs to focus on three core improvements:

Smarter Agent Mode: Ditch the slow, step-by-step clicking. Which can be annoying at times and instead, show a fast visual preview of the Agent's plan and only require user approval for the final action, making automation instant and trustworthy. But I guess it will come to this once we are all familiar with it.

Proactive Memory: The AI shouldn't wait for you to ask. It needs to leverage Browser Memories to nudge you with context—like suggesting related past research or offering to start a task based on the current page you're viewing.

Invisible UI: Make the AI assistant a true co-pilot by having the sidebar vanish when not needed. Improve input flexibility with voice commands or circling page elements to initiate tasks, letting you skip the keyboard whenever possible.

I love it...for months, AI assistants like ChatGPT, or Claude have helped me multiply what I can do, but the experience was always fragmented. You had to jump between apps: copy a text here, summarize there, fix an email somewhere else.

Now, being able to interact inside the browser, across web pages, Gmail, and custom apps, finally makes AI feel like a native layer of the internet, not just another tool.

That said, it’s not perfect yet: sometimes it’s a bit slow, it struggles to read everything on a page, and it can get stuck while switching between tabs. But the potential is massive, especially with the new Agent Mode, which finally lets the model act across tabs and apps instead of being trapped inside one chat window.

This is the first time in a while AI feels truly integrated into everyday workflows, not just smarter, but more "natural".

This looks very helpful! As a Windows user, I'm waiting for that version. Any updates on when it might be available?
And also, is it based on chromium?

This looks very helpful! As a Windows user, I'm waiting for that version. Any updates on when it might be available?

I used Atlas for a repetitive, manual task on a large g-sheet file. The process was slower than expected, and required frequent restarts. For now, performing manually remains faster and more efficient. I hope future updates will make atlas faster and more autonomous. But i know, this is just the starting point so let's hope in better performances for the future ;)

I used atlas for a repetitive task on G-Sheet file. Seem to be slower than expected, and it often freezes and you have to tell it to restart. Right now, manually work looks still faster.
Hoping in new updates!

I used Atlas for a repetitive, manual task on a large Excel file.

The process was slower than expected and required frequent supervision and restarts.

For now, performing the task manually remains faster and more efficient. I hope future updates will make Atlas faster and more autonomous.

I used Atlas for a repetitive, manual task on a large Excel file. The process was slower than expected and required frequent supervision and restarts. For now, performing the task manually remains faster and more efficient. I hope future updates will make Atlas faster, more autonomous, and better suited for handling large-scale repetitive operations.

I wonder if people will actually switch browsers or just use AI inside the ones they already have.