WebMCP is available for early preview • Hacker News Board

Re: WebMCP is available for early preview

#211

Post by _heimdall » Tue, Mar 03, 2026, 3:00 PM UTC

DrScientist wrote: post #209

_heimdall wrote: post #198

Earlier quoted context omitted.

I'm not aware of anything WebMCP could add that wouldn't be more useful as an improvement to accessibility tooling instead. MCP is ultimately another solution to trying to make RPC(ish) situations more RESTful. I.e. they need self-documenting, discoverable APIs. That's exactly what you can get from both HTML and the accessibility tree, though. We don't need another implementation for it. My guess (conjecture here) is…

Isn't Aria there to describe the structure of the page so that say visually impaired users gain the same information as any other user? ie the interpretation of what that page then does, and so the appropriate action to take is largely left to the human user post description - just as web you load a page and look at it - the human brain works out what to do based on those visual and textual clues. This leaves agents…

The website is the API though. HTML is one of the few RESTful systems people still use today, build semantics into the page and humans and LLMs can understand how to use it.

A11y specs and APIs are just a way of presenting those semantics differently, often for those who can't see the page, whether visually impaired or in this case an LLM.

At least in my view, we should expect anything claimed to be artificial intelligence to be able to interact with things much like a human would. I'm not going to build an MCP for a CLI tool, for example, I'll just make sure it has a useful man page or `--help` command.

Re: WebMCP is available for early preview

#212

Post by _heimdall » Tue, Mar 03, 2026, 3:02 PM UTC

charcircuit wrote: post #201

_heimdall wrote: post #179

Earlier quoted context omitted.

That's a huge resource cost though, and simply unnecessary. We should be building semantically valid HTML from the beginning rather than leaning on a GPU cluster to parse the function based on the entire HTML, CSS, and JS on the page (or a screenshot requiring image parsing by a word predictor).

That's the point of solving problems with LLMs. We pay a large resource cost, but in return we get general intelligence to understand things.

We should try to avoid hitting that resource cost on every use where possible though. A CLI tool should have good `--help` docs for example, rather than expecting every inference run to scrub the CLI tool's source code to figure out how to use it.

Re: WebMCP is available for early preview

#213

Post by _heimdall » Tue, Mar 03, 2026, 3:04 PM UTC

charcircuit wrote: post #200

johneth wrote: post #177

Earlier quoted context omitted.

Or just use in the first place and save humans and machines a whole bunch of time.

That's already possible today yet there are still people who don't which is why a more general solution for the screen reader is needed rather than requiring every site developer to do something special.

We shouldn't create general solutions for people building software poorly. We should help people build software better, in this by helping to promote the use of a11y specs.

This is actually exactly where model providers could be doing some good. If they said a11y is the way for LLMs to interact with the web and helped push developers to docs, tutorials, etc the web would be better off. Google did effectively just that with HTTPS, they told everyone use it or lose SEO value rather than slapping some solution on Google's end to paper over poor security practices.

Re: WebMCP is available for early preview

#214

Post by sarkarsh » Tue, Mar 03, 2026, 7:18 PM UTC

manveerc wrote: post #8

BeefySwain wrote: post #3

Can someone explain what the hell is going on here? Do websites want to prevent automated tooling, as indicated by everyone putting everything behind Cloudfare and CAPTCHAs since forever, or do websites want you to be able to automate things? Because I don't see how you can have both. If I'm using Selenium it's a problem, but if I'm using Claude it's fine??

In my opinion sites that want agent access should expose server-side MCP, server owns the tools, no browser middleman. Already works today. Sites that don’t want it will keep blocking. WebMCP doesn’t change that. Your point about selenium is absolutely right. WebMCP is an unnecessary standard. Same developer effort as server-side MCP but routed through the browser, creating a copy that drifts from the actual UI. For…

[dead]

Re: WebMCP is available for early preview

#215

Post by jauntywundrkind » Tue, Mar 03, 2026, 7:21 PM UTC

arjunchint wrote: post #208

jauntywundrkind wrote: post #205

Earlier quoted context omitted.

I just really dislike the uselessness of people who naysay & dont engage! This poor world suffers SO MUCH from Brandolini's Law, from bad information being so easy to create. My heart is torn by bad engagement, by misdirection, away from the good and the interesting and the possible, and there's such an asymmetry that the truth and possibility face, so many ways for potential to be sapped and drained. Hackers deserve…

There have been thousands-millions of proposals since the dawn of the internet that got nowhere. To exist is to recognize the material constraints of reality; there are things humans won't ever discover. Ergo we have to prioritize what is useful. This proposal is not useful. It goes against the fundamental interest of website owners to differentiate and build up a moat around direct user relationship and data. WebMCP…

Said as if it's all just happening around us, some hand of fate moving things to where they will be. I am spiritually opposed to this point of view, and find it dead and against the hacker spirit. We are participants in the marketplace of ideas. Our words here should be used to inform and steer ourselves & each other.

If you want to not pay attention, go right ahead! But kindly don't waste everyone's time by commenting with your irregard & wasting all of our time.

This proposal is very useful. It goes for the users to help them disentangle themselves from the heinous moats of control that software fiefs try to erect around themselves. Google is helping companies that want to better relate to their users, to give those users much amplified agency to access the software and systems on their own terms.

That's how software should win! That should be a colossal crushing victory over the control & manipulation that the human spirit detests and loaths.

That is a huge win. Doomsaying is evil and bad, and embarrassing to hacker-kind. Breath life into better futures. At least consider with curiosity and interest the possibilities of better: so often the future is found by those who do look to the past, and see what was passed over. Be the engaged. Be the interested. Seek value & interest.

Re: WebMCP is available for early preview

#216

Post by cindyllm » Tue, Mar 03, 2026, 7:33 PM UTC

jauntywundrkind wrote: post #215

arjunchint wrote: post #208

Earlier quoted context omitted.

There have been thousands-millions of proposals since the dawn of the internet that got nowhere. To exist is to recognize the material constraints of reality; there are things humans won't ever discover. Ergo we have to prioritize what is useful. This proposal is not useful. It goes against the fundamental interest of website owners to differentiate and build up a moat around direct user relationship and data. WebMCP…

Said as if it's all just happening around us, some hand of fate moving things to where they will be. I am spiritually opposed to this point of view, and find it dead and against the hacker spirit. We are participants in the marketplace of ideas. Our words here should be used to inform and steer ourselves & each other. If you want to not pay attention, go right ahead! But kindly don't waste everyone's time by commenti…

[dead]

Re: WebMCP is available for early preview

#217

Post by otterley » Wed, Mar 04, 2026, 12:21 AM UTC

jauntywundrkind wrote: post #215

arjunchint wrote: post #208

Earlier quoted context omitted.

There have been thousands-millions of proposals since the dawn of the internet that got nowhere. To exist is to recognize the material constraints of reality; there are things humans won't ever discover. Ergo we have to prioritize what is useful. This proposal is not useful. It goes against the fundamental interest of website owners to differentiate and build up a moat around direct user relationship and data. WebMCP…

Said as if it's all just happening around us, some hand of fate moving things to where they will be. I am spiritually opposed to this point of view, and find it dead and against the hacker spirit. We are participants in the marketplace of ideas. Our words here should be used to inform and steer ourselves & each other. If you want to not pay attention, go right ahead! But kindly don't waste everyone's time by commenti…

> But kindly don't waste everyone's time by commenting with your irregard & wasting all of our time.

Don't tell people not to disagree with you. The "marketplace of ideas" you celebrate here is full of disagreement.

If you disagree with people, just make the best substantive argument you can. Don't characterize your fellow participants. Criticize the idea, not the person. And expect to lose your fair share of arguments.

Re: WebMCP is available for early preview

#218

Post by dakolli » Wed, Mar 04, 2026, 2:48 AM UTC

zamadatix wrote: post #168

dakolli wrote: post #120

Is this just devtools protocol wrapped by an MCP? I've been doing this with go-rod for two years... https://github.com/go-rod/rod

Rod seems to be about automating the local browser itself via MCP, through which you can try to self-automate websites loaded in the browser. Interestingly, it seems Google nowadays has its own official implementation of this https://github.com/ChromeDevTools/chrome-devtools-mcp WebMCP seems to be about the authors of websites being able to publish a list of custom built-in tools the page has available for LLM agents…

Thanks for this, I had no clue DevTools had an MCP, could be useful.

Re: WebMCP is available for early preview

#219

Post by jauntywundrkind » Wed, Mar 04, 2026, 6:09 AM UTC

otterley wrote: post #217

jauntywundrkind wrote: post #215

Earlier quoted context omitted.

Said as if it's all just happening around us, some hand of fate moving things to where they will be. I am spiritually opposed to this point of view, and find it dead and against the hacker spirit. We are participants in the marketplace of ideas. Our words here should be used to inform and steer ourselves & each other. If you want to not pay attention, go right ahead! But kindly don't waste everyone's time by commenti…

> But kindly don't waste everyone's time by commenting with your irregard & wasting all of our time. Don't tell people not to disagree with you. The "marketplace of ideas" you celebrate here is full of disagreement. If you disagree with people, just make the best substantive argument you can. Don't characterize your fellow participants. Criticize the idea, not the person. And expect to lose your fair share of argumen…

I LOVE DISAGREEMENT.

But I expect some actual engagement, & willingness to explore the topic. I maintain that you have avoided engagement, and blanketly worked to shut down thought and discussion, rather than explore ideas.

> If you disagree with people, just make the best substantive argument you can.

I've done just that, raising and supporting my points and responding to your arguments. But you ignore the previous discussion and make new posts that don't seem to ever build off anything that's happened so far or make any acknowledgement of past discussion: it's always some new point, or trying to score some technical win over me, chastizing me for form, as you do as here. None of it builds.

As I see it, you've been wasting everyone's time, and since you don't show any signs of considering how this might be different, and since you won't engage in points raised. I regrettably think it's better to share my opinion that we all would be better off giving you no attention at all, if you keep posting without participating as you have done.

> Don't characterize your fellow participants. Criticize the idea, not the person.

You deserve criticism for the way you have misdirected away from exploration or consideration, in a non-engaged fashion. It has been a mis-service to the topic, and a mis-service to everyone else's time.

This is getting less and less specific & less to the point from you. You seem to have tons of time to write these letters, but no time to engage in the topic. I don't know what it is you are after, but being policed by you like this is just further distracting from anything of relevance. Don't do this. It's continuing to be a negative drain.

I think it's incredible that we have this collaborative system where people and LLMs can work together. You are free to disagree, but so far, you haven't, you haven't shown any recongition at all that you understand how this might be different, and it seems like you've just been monologuing your naysaying over-top the points I have raised. Without showing signs that you've read or heard anything said.

Re: WebMCP is available for early preview

#220

Post by DrScientist » Wed, Mar 04, 2026, 10:12 AM UTC

_heimdall wrote: post #211

DrScientist wrote: post #209

Earlier quoted context omitted.

Isn't Aria there to describe the structure of the page so that say visually impaired users gain the same information as any other user? ie the interpretation of what that page then does, and so the appropriate action to take is largely left to the human user post description - just as web you load a page and look at it - the human brain works out what to do based on those visual and textual clues. This leaves agents…

The website is the API though. HTML is one of the few RESTful systems people still use today, build semantics into the page and humans and LLMs can understand how to use it. A11y specs and APIs are just a way of presenting those semantics differently, often for those who can't see the page, whether visually impaired or in this case an LLM. At least in my view, we should expect anything claimed to be artificial intell…

I think you are confusing two things.

- the semantics of a form and a button and the resulting http POST/GET - and what the page actually does!

So I can have two pages - both with html forms - what they actually do on submission might be completely different - one buys a potted plant the other submits a tax return.

ie the meaning of the action is in the non-semantic elements - the free text, the images, the context.

This is the stuff that's hard for the agent to easily determine - is this a form for submitting a tax return or not?

If what you said is true then there would be already agents out there that use ARIA info to seemlessly operate the web. As far as I can see people have tried to use that information to improve agents use of the web - but have met limited success - and that's for well annotated sites - not because sites aren't ARIA enabled.