BrowserOS is the best browser I've used so far

Ironhead

Junior Member
Joined
Apr 10, 2023
Messages
111
Reaction score
68
I haven't used comet browser (by perplexity) as I mainly use linux but browserOS is great, I can ask it to do things like research on my feed, reply to people, like stuff. I've set up a few built in agents that have 15-20 step processes that work flawlessly with gemini 2.5 flash API.

You could get real creative with these, it even acts as an MCP server meaning tools like Claude Code (MCP clients) can control it for you.

Couple other features that make it useful:
  • Prompt -> Agent
  • Record screen and narrate -> Agent
It would make life a lot easier if dolphin anty, adspower etc included browserOS tech

Screenshot_20251019_173939.png
 
This is looking great. You can actually train accounts to have a personality instead of just handling actions like robots.
 
Hey It looks great!
I will try it this week.
Great share mate!
 
checking this out today. have seen other takes on ai in browser now, but never heard of this before. lots of cool new ways to automate browser actions!
 
Would love a more detailed step-by-step use.
 
I've never heard of it but it looks greater than great, I will try it out
 
Last edited:
Also openai just launched chatgpt atlas, an agentic browser, but only available on Mac for now
 
This things amazing I told it to create me a random gmail account and it did that 99% I told it to browse autotrader to find a specific car it did that aswell 100%
 
This things amazing I told it to create me a random gmail account and it did that 99% I told it to browse autotrader to find a specific car it did that aswell 100%
Yep, and as better models come out for browser control it will get more accurate, and faster.

There are benchmarks/leaderboards for browser use and computer use agents too.
 
I haven't used comet browser (by perplexity) as I mainly use linux but browserOS is great, I can ask it to do things like research on my feed, reply to people, like stuff. I've set up a few built in agents that have 15-20 step processes that work flawlessly with gemini 2.5 flash API.

You could get real creative with these, it even acts as an MCP server meaning tools like Claude Code (MCP clients) can control it for you.

Couple other features that make it useful:
  • Prompt -> Agent
  • Record screen and narrate -> Agent
It would make life a lot easier if dolphin anty, adspower etc included browserOS tech

View attachment 481328
This looks great! Thanks for the share!
 
interesting find. i hadn’t heard of BrowserOS before but from what you say it sounds solid.


quick thoughts: since it’s new you’ll want to test it with your current workflow—check extension compatibility, make sure your proxies/device setup works fine and see how it performs under real traffic. also keep a backup browser ready just in case.
 
What API and model do you use to interact with the browser?
 
Back
Top