Earlier quoted context omitted.
> building a more AI friendly DOM You might consider the Accessibility Tree and its semantics. Plain divs are basically filtered out so you're left with interactive objects and some structural/layout cues.
I've been trying (albeit not very hard) to build an accessibility library and toolset that can be exposed via mcp server. I think it has the potential to be much more ergonomic for generalized computer-use agents than stuff like playwright or the classic screenshot approach. Low latency computer use is another thing that I'd like to solve. The issue is mac and windows accessibility APIs are opaque and I have no idea…
Show HN: Nxtscape – an open-source agentic browser
211–220 of 228 posts
Re: Show HN: Nxtscape – an open-source agentic browser
#212Earlier quoted context omitted.
I've been trying (albeit not very hard) to build an accessibility library and toolset that can be exposed via mcp server. I think it has the potential to be much more ergonomic for generalized computer-use agents than stuff like playwright or the classic screenshot approach. Low latency computer use is another thing that I'd like to solve. The issue is mac and windows accessibility APIs are opaque and I have no idea…
Neat, is it on github?
I will try to publish something on gh this weekend.
Re: Show HN: Nxtscape – an open-source agentic browser
#2131. How will you earn money? Sustainig the project? 2. Which LLM is used? Only "local LLM" is mentioned. Hardware requirements?
Reg. first part, our thinking right now is similar to other OSS projects, we will have enterprise offering and charge for that.
Re: Show HN: Nxtscape – an open-source agentic browser
#214Earlier quoted context omitted.
Yes it would matter. The AI might be I in your eyes, but it is still A.
So is Chrome. Very artificial. It's still not a robot for the purposes of robots.txt. What coherent definition of robot excludes Chrome but includes this?
No meatsack in the loop making decisions and pushing the button? Robots.txt applies.
Re: Show HN: Nxtscape – an open-source agentic browser
#215https://chromium.googlesource.com/chromium/src/+/main/docs/l...
It seems like gemini 2.5 is pre-trained on chromium given it has a lot of context into the current state of the codebase. That may help a bit.
Also, I've not tried this yet, but maybe using an MCP server over a clangd language server is a solution to the context problem?
https://chromium.googlesource.com/chromium/src/+/master/docs...
Re: Show HN: Nxtscape – an open-source agentic browser
#216In fact, we built one, rtrvr.ai that has even better Web Agent performance than Open AI's Operator with human assistant and 7x faster than leading competitor: https://www.rtrvr.ai/blog/web-bench-results
Your Accessibility Tree requirement is a poor excuse, rather you should build up an agent from a first principles understanding of DOM interactions.
A browser is a SERIOUS security risk, you need a dedicated team to just pull in the latest security patches that Google pushes to Chromium or your users are sitting ducks to exploits and hacks...
Re: Show HN: Nxtscape – an open-source agentic browser
#217IMO comments so far seem to be not seeing the forest for the trees -- I can imagine incredible value for myself in a browser that hooks into a local LLM, writes everything it sees to a local timestamped database (oversimplification), parses and summarizes everything you interact with (again, oversimplification -- this would be tunable and scriptable), exposes Puppeteer-like functionality that is both scriptable via c…
This is basically what Microsoft wants to do with Recall and they got slammed for it. Which drives me nuts because it's the only feature from the recent AI hype wave that excites me because it's the only thing so far that sounds like it will actually make my life better. But then I thought about it a bit more and I realized that what I really want is not AI, I just want my computer to have a detailed local history an…
Sheesh.
As someone who uses web browsers that delete all session data when they're closed, and routinely wipes any "recently used" lists and temporary files in all operating systems I use, the thought of the machines I'm using remembering my usage behavior to such an extent is terrifying to me.
I mean, I get why those features would be appealing. I just have zero trust in the companies that build such software, because they've violated my trust time and time again. Yet I'm expected to suddenly trust them in this case? Not a chance. Not when the data that I would be entrusting them with for the features you mention are a literal gold mine for them. Trusting a trillion-dollar corporation with a history of privacy violations and hostility towards its users is just unthinkable for me, no matter what features I might be missing out on.
It's unfortunate that computing has come to this, but I choose the hardware and software I use very carefully. In cases where I'm forced to use a system I don't trust, I try my best to limit my activity and minimize the digital footprints I leave behind. I prefer using open source software for this reason, but even that is carefully selected, since OSS is easily corruptible by unscrupulous developers, and companies that use it as a marketing tactic.
The only way I might use software with that level of intrusion is if I've inspected every line of it, I run it inside a container with very limited permissions, or if I've written it myself. Needless to say, it's more likely I'll just miss out on using those features instead, and I'm fine with that.
Re: Show HN: Nxtscape – an open-source agentic browser
#218Re: Show HN: Nxtscape – an open-source agentic browser
#2191. How will you earn money? Sustainig the project? 2. Which LLM is used? Only "local LLM" is mentioned. Hardware requirements?
for local LLM, you can download ollama and you can download 4B or 8B variants. Reg. first part, our thinking right now is similar to other OSS projects, we will have enterprise offering and charge for that.
Re: Show HN: Nxtscape – an open-source agentic browser
#220Earlier quoted context omitted.
Like hundreds of millions of people want to chat with a computer. At least.
This is simply not true in the productivity space (which is the context here). People just want to get shit done.