I’ve been wanting to use browser automation to track private twitter accounts that I already follow, since the API obviously won't cut it. My main concern with this browser's automation features is account safety—how well does it mask itself? I'm worried it might trigger the platform's anti-bot detection and get my account flagged or banned :(
Just tried it out, the Space concept is exactly what I’ve wanted all along,which so straightforward: agent gets its own independent space but still holds my credentials. Browsers like Dia, which evolved from Arc (working on the workspace concept very early on), should have taken this route ages ago. What a missed opportunity that they didn't integrate it this way.
This might be a niche use case, but I love using browser agents for web games.
A browser that can run multiple agent-controlled sessions at the same time basically turns multiboxing from a chore into a one-click experience. My productivity may not improve, but my game progression definitely will.
smart move, when I was using agent browser,I have to watch it execute commands one by one, it’s really inefficient. using JS code to execute all commands at once is smart.
I have read the repo, from what I can see is the thing used to connect the browser is writen in js, command and tool all stay in js makes the execution faster, make sense, but I m not sure, coz I haven’t read about their benchmark, could just a story lol
I’ve been wanting to use browser automation to track private twitter accounts that I already follow, since the API obviously won't cut it. My main concern with this browser's automation features is account safety—how well does it mask itself? I'm worried it might trigger the platform's anti-bot detection and get my account flagged or banned :(
Just tried it out, the Space concept is exactly what I’ve wanted all along,which so straightforward: agent gets its own independent space but still holds my credentials. Browsers like Dia, which evolved from Arc (working on the workspace concept very early on), should have taken this route ages ago. What a missed opportunity that they didn't integrate it this way.
This might be a niche use case, but I love using browser agents for web games.
A browser that can run multiple agent-controlled sessions at the same time basically turns multiboxing from a chore into a one-click experience. My productivity may not improve, but my game progression definitely will.
smart move, when I was using agent browser,I have to watch it execute commands one by one, it’s really inefficient. using JS code to execute all commands at once is smart.
Why javascript ?
I have read the repo, from what I can see is the thing used to connect the browser is writen in js, command and tool all stay in js makes the execution faster, make sense, but I m not sure, coz I haven’t read about their benchmark, could just a story lol
Thanks bro, I'll check the repo by myself.
My point is here: There are many popular AI-Friendly PL, eg. Python, Rust ...
I also see many CDP-based framework which implemented by Python-Binding libs, They works very well.
I think javascript is not the only option.