Live data from Hacker News

Show HN: Nxtscape – an open-source agentic browser

github.com

101–110 of 228 posts

Re: Show HN: Nxtscape – an open-source agentic browser

#103
post #101

This is cool thanks for sharing What is the tech around the thing that segments out DOM elements automatically and shows the visual representation. I think something like this would be great for automated UI testing agents?

Oh that's from browser use buildDOMTree.js. check it out on their github.

Re: Show HN: Nxtscape – an open-source agentic browser

#104
Suffers from the same problem as all other AI "workflows"--no one wants to fucking chat with a computer (Brave sort-of does this, and it's god-awful). A chat interface should only be used as a fallback if the agent is too dumb to figure out what I want.

A chat interface works for ChatGPT because most folks use it as a pseudo-search, but productivity tools are (broadly speaking) not generative, therefore shouldn't be using freeform inputs. I have many thoughts on fixing this, and it's a very hard problem, but simply slapping an LLM onto Chrome is just lazy. I don't mean to be overly negative, but it's kind of wild to see YC funding slop like this.

And that's exactly what this is: slop. There's no technical creativity here, this isn't a new product segment, it barely deserves the "hey bro, this might be a feature, not a product" startup 101 criticism. It's what ChatGPT would spit out if you asked it what a good startup idea would be in 2025. All we need to do, even if we were being as charitable as possible, is ask who's doing the heavy lifting here (hint: it's not in the Github repo).

Re: Show HN: Nxtscape – an open-source agentic browser

#105
post #100

Why obfuscate the LLM system prompt in your Github repo when it's going to be completely visible in the network inspector?

Not on purpose, just got compressed during production build through webpack. will get that fixed.

Thank you for the clarification!

While reviewing the prompt's capabilities, I had an idea: implementing a Greasemonkey/Userscript-style system, where users could inject custom JavaScript or prompts based on URLs, could be a powerful way to enhance website interactions.

For instance, consider a banking website with a cumbersome data export process that requires extra steps to make the data usable. Imagine being able to add a custom button to their UI (or define a custom MCP function) specifically for that URL, which could automatically pull and format the data into a more convenient format for plain text accounting.

Re: Show HN: Nxtscape – an open-source agentic browser

#106

This is definitely a winners take all market. Kudos for giving it a shot, but imo browser projects are just too big for a team of 2/3. Plus, google has already demoed at IO the first hint at this. IMO you just cannot move fast enough to grab enough market share as a first/second mover before google just does it on chrome, and that's assuming you can outcompete with Dia in the first place. Even browser-use can do this…

Thanks for the honest feedback!

Definitely agree there is good amount of competition here.

But we do think there is a gap in the market for open-source, community driven and privacy-first AI browser. (Something like Brave?)

Re: Show HN: Nxtscape – an open-source agentic browser

#109

Earlier quoted context omitted.

Not to pull a "why should I use Dropbox when I have rsync" but why should we use this over adding a Playwright MCP to Claude Desktop or similar? Does having access to Chromium internals give you any super powers over connecting over the Chrome Devtools Protocol?

I would take the position of "why use this when I have eyes and hands and a brain?"

My guess is that this is for impatient people; people who think that the prescribed use cases are somehow necessary for their "workflows"; people who subscribe to terms like "cognitive friction" within the context of these use cases; people who are...sort of lazy.

Re: Show HN: Nxtscape – an open-source agentic browser

#110

Earlier quoted context omitted.

What do you mean? This AI cannot scrape multiple links automatically? Like "make a summary of all the recipes linked in this page" kind of stuff? If it can it definitely meets the definition of scraping.

I think what he means is it is not just generally crawling and scraping, and uses a more targeted approach. Equivalent to a user going to each of those sites, just more efficiently.

I'm guessing that would ideally mean only reading the content the user would otherwise have gone through. I wonder if that's the case and if it's guaranteed.

Maybe some new standards and maybe a user configurable per site permissions may make it better?

I'm curious to see how this will turn out to be.

Post reply on HN