TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Show HN: Browser MCP – Automate your browser using Cursor, Claude, VS Code

616 点作者 namukang大约 1 个月前

51 条评论

rmac大约 1 个月前
[!warning!]<p>1) this projects&#x27; chrome extension sends detailed telemetry to posthog and amplitude:<p>- <a href="https:&#x2F;&#x2F;storage.googleapis.com&#x2F;cobrowser-images&#x2F;telemetry.png" rel="nofollow">https:&#x2F;&#x2F;storage.googleapis.com&#x2F;cobrowser-images&#x2F;telemetry.pn...</a><p>- <a href="https:&#x2F;&#x2F;storage.googleapis.com&#x2F;cobrowser-images&#x2F;pings.png" rel="nofollow">https:&#x2F;&#x2F;storage.googleapis.com&#x2F;cobrowser-images&#x2F;pings.png</a><p>2) this project includes source for the local mcp server, but not for its chrome extension, which is likely bundling <a href="https:&#x2F;&#x2F;github.com&#x2F;ruifigueira&#x2F;playwright-crx">https:&#x2F;&#x2F;github.com&#x2F;ruifigueira&#x2F;playwright-crx</a> without attribution<p>super suss
评论 #43624154 未加载
评论 #43623145 未加载
评论 #43620709 未加载
bhouston大约 1 个月前
So the website claims:<p>&quot;Avoids bot detection and CAPTCHAs by using your real browser fingerprint.&quot;<p>Yeah, not really.<p>I&#x27;ve used a similar system a few weeks back (one I wrote myself), having AI control my browser using my logged in session, and I started to get Captcha&#x27;s during my human sessions in the browser and eventually I got blocked from a bunch of websites. Now that I&#x27;ve stopped using my browser session in that way, the blocks eventually went away, but be warned, you&#x27;ll lose access yourself to websites doing this, it isn&#x27;t a silver bullet.
评论 #43614754 未加载
评论 #43624297 未加载
评论 #43614578 未加载
评论 #43615784 未加载
评论 #43619122 未加载
StevenNunez大约 1 个月前
I feel like I slept for a day and now MCPs are everywhere... I don&#x27;t know what MCPs are and at this point I&#x27;m too afraid to ask.
评论 #43616648 未加载
评论 #43616347 未加载
评论 #43617729 未加载
评论 #43617292 未加载
评论 #43620074 未加载
andy_ppp大约 1 个月前
When I go to a shopping website I want to be able to tell my browser &quot;hey please go through all the sideboards on this list and filter out for the ones that are larger than 155cm and smaller than 100cm, prioritise the ones with dark wood and space for vinyl records which are 31.43cm tall&quot; for example.<p>Is there any browser that can do this yet as it seems extremely useful to be able to extract details from the page!
评论 #43616303 未加载
评论 #43655816 未加载
评论 #43624324 未加载
评论 #43616351 未加载
neilellis大约 1 个月前
Well done, just tested on Claude Desktop and it worked smoothly and a lot less clunky than playwright. This is the right direction to go in.<p>I don&#x27;t know if you&#x27;ve done it already, but it would be great to pause automation when you detect a captcha on the page and then notify the user that the automation needs attention. Playwright keeps trying to plough through captchas.
thenaturalist大约 1 个月前
Crazy, in looking up some info on the web and creating a Spreadsheet on Google Sheets to insert the results, it worked almost perfectly the first time and completely failed subsequently on 8-10 different tries.<p>Is there an issue with the lag between what is happening in the browser and the MCP app (in my case Claude Desktop)?<p>I have a feeling the first time I tried it, I was fast enough clicking the &quot;Allow for this chat&quot; permissions, whereas by the time I clicked the permission on subsequent chats, the LLM just reports &quot;It seems we had an issue with the click. Let me try again with a different reference.&quot;.<p>Actions which worked flawlessly the first time (rename a Google spreadsheet by clicking on the title and inputting the name) fail 100% of subsequent attempts.<p>Same with identifying cells A1, B1, etc. and inserting into the rows.<p>Almost perfect on 1st try, not reproducible in 100% of attempts afterwards.<p>Kudos to how smooth this experience is though, very nice setup &amp; execution!<p>EDIT 2: The lag &amp; speed to click the allow action make it seemingly unusable in Claude Desktop. :(
评论 #43615351 未加载
评论 #43648672 未加载
评论 #43617177 未加载
评论 #43622009 未加载
nonethewiser大约 1 个月前
Stuff like this makes me giddy for manual tasks like reimbursement requests. Its such a chore (and it doesnt help our process isnt great).<p>Every month, go to service providers, log in, find and download statement, create google doc with details filled in, download it, write new email and upload all the files. Maybe double chek the attachments are right but that requires downloading them again instead of being able to view in email).<p>Automating this is already possible (and a real expense tracking app can eliminate about half of this work) but I think AI tools have the potential to elminate a lot of the nittier-grittier specification of it. This is especially important because these sorts of workflows are often subject to little changes.
doug_life大约 1 个月前
This may be obvious to most here, but you need Node.js installed for the MCP server to run. This critical detail is not in the set up instructions.
评论 #43617638 未加载
评论 #43617639 未加载
serverlessmania大约 1 个月前
Did something similar but controls a hardware synth, allowing me to do sound design without touching the physical knobs: <a href="https:&#x2F;&#x2F;github.com&#x2F;zerubeus&#x2F;elektron-mcp">https:&#x2F;&#x2F;github.com&#x2F;zerubeus&#x2F;elektron-mcp</a>
评论 #43614207 未加载
Gehinnn大约 1 个月前
Would be nice if it could use the Accessibility Tree from chrome dev tools to navigate the page instead of relying on screenshots (<a href="https:&#x2F;&#x2F;developer.chrome.com&#x2F;blog&#x2F;full-accessibility-tree" rel="nofollow">https:&#x2F;&#x2F;developer.chrome.com&#x2F;blog&#x2F;full-accessibility-tree</a>)
评论 #43617312 未加载
amendegree大约 1 个月前
So is MCP the new RPA (Robotics Process Automation)? Like generic yahoo pipes?
评论 #43614150 未加载
评论 #43620153 未加载
评论 #43614117 未加载
cadence-大约 1 个月前
Doesn&#x27;t work on Windows:<p>2025-04-07T18:43:26.537Z [browsermcp] [info] Initializing server... 2025-04-07T18:43:26.603Z [browsermcp] [info] Server started and connected successfully 2025-04-07T18:43:26.610Z [browsermcp] [info] Message from client: {&quot;method&quot;:&quot;initialize&quot;,&quot;params&quot;:{&quot;protocolVersion&quot;:&quot;2024-11-05&quot;,&quot;capabilities&quot;:{},&quot;clientInfo&quot;:{&quot;name&quot;:&quot;claude-ai&quot;,&quot;version&quot;:&quot;0.1.0&quot;}},&quot;jsonrpc&quot;:&quot;2.0&quot;,&quot;id&quot;:0} node:internal&#x2F;errors:983 const err = new Error(message); ^<p>Error: Command failed: FOR &#x2F;F &quot;tokens=5&quot; %a in (&#x27;netstat -ano ^| findstr :9009&#x27;) do taskkill &#x2F;F &#x2F;PID %a at genericNodeError (node:internal&#x2F;errors:983:15) at wrappedFn (node:internal&#x2F;errors:537:14) at checkExecSyncError (node:child_process:882:11) at execSync (node:child_process:954:15)
评论 #43614684 未加载
评论 #43614683 未加载
josefrichter大约 1 个月前
What I used this for:<p>&quot;Go to <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;upvoted?id=josefrichter">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;upvoted?id=josefrichter</a>, summarize what topics I am interested in, and then from the homepage pick articles I might be interested in.&quot;<p>Works like a charm.
washedDeveloper大约 1 个月前
Can you add a license to your code along with open sourcing the chrome extension?
makingstuffs大约 1 个月前
I don&#x27;t see how an MCP can be useful for browsing the net and doing things like shopping as has been suggested. Large companies such as CloudFlare have spent millions on, and made a business from, bot detection and blocking.<p>Do we suppose they will just create a backdoor to allow _some_ bots in? If they do that how long will it be before other bots impersonate them? It seems like a bit of a fad from my small mind.<p>Suppose it does become a thing, what then? We end up with an internet which is heavily optimised for bots (arguably it already is to an extent) and unusable for humans?<p>Wild.
评论 #43619025 未加载
评论 #43619180 未加载
评论 #43620557 未加载
评论 #43620056 未加载
hliyan大约 1 个月前
Ideally, shouldn&#x27;t this be the native experience of most &quot;sites&quot; on the internet? We&#x27;ve built an entire user experience around serving users rich, two dimensional visual content that is not machine-readable and are now building a natural language command line layer on top of it. Why not get rid of the middleware and present users a direct natural language interface to the application layer?
buttofthejoke大约 1 个月前
Why use this over Puppeteer or Playwright extensions?
评论 #43613905 未加载
Fernicia大约 1 个月前
Any plans to make a Firefox version?
评论 #43614105 未加载
DebtDeflation大约 1 个月前
In the Task Automation demo, how does it know all of the attributes of the motorcycle he is trying to sell? Is it relying on the underlying LLM&#x27;s embedded knowledge? But then how would it know the price and mileage? Is there some underlying document not referenced in the demo? Because that information is not in the prompt.
pavelfeldman大约 1 个月前
I mean no disrespect, but this looks like an outdated clone of <a href="https:&#x2F;&#x2F;github.com&#x2F;microsoft&#x2F;playwright-mcp">https:&#x2F;&#x2F;github.com&#x2F;microsoft&#x2F;playwright-mcp</a><p><a href="https:&#x2F;&#x2F;github.com&#x2F;microsoft&#x2F;playwright-mcp&#x2F;blob&#x2F;main&#x2F;src&#x2F;tools&#x2F;tool.ts">https:&#x2F;&#x2F;github.com&#x2F;microsoft&#x2F;playwright-mcp&#x2F;blob&#x2F;main&#x2F;src&#x2F;to...</a> <a href="https:&#x2F;&#x2F;github.com&#x2F;BrowserMCP&#x2F;mcp&#x2F;blob&#x2F;main&#x2F;src&#x2F;tools&#x2F;tool.ts">https:&#x2F;&#x2F;github.com&#x2F;BrowserMCP&#x2F;mcp&#x2F;blob&#x2F;main&#x2F;src&#x2F;tools&#x2F;tool.t...</a>
评论 #43617583 未加载
评论 #43617461 未加载
icelancer大约 1 个月前
I just run into a bunch of errors on my Windows machine + Chrome when connected over remote-ssh. Extension installed, tab enabled, npx updated&#x2F;installed, etc.<p>2025-04-07 10:57:11.606 [info] rmcp: Starting new stdio process with command: npx @browsermcp&#x2F;mcp@latest<p>2025-04-07 10:57:11.606 [error] rmcp: Client error for command spawn npx ENOENT<p>2025-04-07 10:57:11.606 [error] rmcp: Error in MCP: spawn npx ENOENT<p>2025-04-07 10:57:11.606 [info] rmcp: Client closed for command<p>2025-04-07 10:57:11.606 [error] rmcp: Error in MCP: Client closed<p>2025-04-07 10:57:11.606 [info] rmcp: Handling ListOfferings action<p>2025-04-07 10:57:11.606 [error] rmcp: No server info found<p>---<p>EDIT: Ended up fixing it by patching index.js. killProcessOnPort() was the problem. Can hit me up if you have questions, I cannot figure out how to put readable code in HN after all these years with the fake markdown syntax they use.
评论 #43614482 未加载
评论 #43614446 未加载
sdotdev大约 1 个月前
Still slightly confused on what MCPs are but looking at this it does look useful
评论 #43618248 未加载
评论 #43617912 未加载
BrandiATMuhkuh大约 1 个月前
This is really well done! Very cool.<p>I wonder if it&#x27;s possible to add such plugins to election apps (e.g.: Slack). It would be such a nice experience if I could just connect my AI of choice to a local app.
评论 #43614370 未加载
评论 #43614408 未加载
wifipunk大约 1 个月前
Setting this up for claude desktop and cursor was alright. Works well out of the box with little setup, and I like that it attached to my active browser tab. Keep up the good work.
qwertox大约 1 个月前
MCP seems to be JavaScript&#x27;s trojan horse into AI.
评论 #43614473 未加载
评论 #43614688 未加载
otherayden大约 1 个月前
I literally started working on the same exact idea last night haha. Great work OP. I&#x27;m curious, how are you feeding the web data to the LLM? Are you just passing the entire page contents to it and then having it interact with the page based on CSS selectors&#x2F;xpath? Also, what are your thoughts on letting it do its own scripting to automate certain tasks?
metadat大约 1 个月前
Bot Detection Evasion is becoming an increasingly relevant topic. Even for non-abusive automation, it&#x27;s now a necessary consideration.<p>Interesting research and reading via the HN search portal: <a href="https:&#x2F;&#x2F;hn.algolia.com&#x2F;?q=bot+detection" rel="nofollow">https:&#x2F;&#x2F;hn.algolia.com&#x2F;?q=bot+detection</a>
behnamoh大约 1 个月前
What I don&#x27;t like about LLMs is that people keep re-inventing the wheel over and over. For example, we&#x27;ve been able to control browsers using GPT for about 2 years now:<p>- <a href="https:&#x2F;&#x2F;github.com&#x2F;mayt&#x2F;BrowserGPT">https:&#x2F;&#x2F;github.com&#x2F;mayt&#x2F;BrowserGPT</a><p>- <a href="https:&#x2F;&#x2F;github.com&#x2F;TaxyAI&#x2F;browser-extension">https:&#x2F;&#x2F;github.com&#x2F;TaxyAI&#x2F;browser-extension</a><p>- <a href="https:&#x2F;&#x2F;github.com&#x2F;browser-use&#x2F;browser-use">https:&#x2F;&#x2F;github.com&#x2F;browser-use&#x2F;browser-use</a><p>- <a href="https:&#x2F;&#x2F;github.com&#x2F;Skyvern-AI&#x2F;skyvern">https:&#x2F;&#x2F;github.com&#x2F;Skyvern-AI&#x2F;skyvern</a><p>- <a href="https:&#x2F;&#x2F;github.com&#x2F;m1guelpf&#x2F;browser-agent">https:&#x2F;&#x2F;github.com&#x2F;m1guelpf&#x2F;browser-agent</a><p>- <a href="https:&#x2F;&#x2F;github.com&#x2F;richardyc&#x2F;Chrome-GPT">https:&#x2F;&#x2F;github.com&#x2F;richardyc&#x2F;Chrome-GPT</a><p>- <a href="https:&#x2F;&#x2F;github.com&#x2F;handrew&#x2F;browserpilot">https:&#x2F;&#x2F;github.com&#x2F;handrew&#x2F;browserpilot</a><p>- <a href="https:&#x2F;&#x2F;github.com&#x2F;ishan0102&#x2F;vimGPT">https:&#x2F;&#x2F;github.com&#x2F;ishan0102&#x2F;vimGPT</a><p>- <a href="https:&#x2F;&#x2F;github.com&#x2F;Jiayi-Pan&#x2F;GPT-V-on-Web">https:&#x2F;&#x2F;github.com&#x2F;Jiayi-Pan&#x2F;GPT-V-on-Web</a>
评论 #43614367 未加载
评论 #43619278 未加载
评论 #43618555 未加载
评论 #43620724 未加载
评论 #43616750 未加载
webprofusion大约 1 个月前
Or just use Playwright MCP: <a href="https:&#x2F;&#x2F;github.com&#x2F;microsoft&#x2F;playwright-mcp">https:&#x2F;&#x2F;github.com&#x2F;microsoft&#x2F;playwright-mcp</a>
rahimnathwani大约 1 个月前
This is cool. I&#x27;m curious why you chose to use an extension, rather than getting the user to run Chrome with remote debugging turned on?
评论 #43613934 未加载
评论 #43613912 未加载
101008大约 1 个月前
Good, just what we needed. More bots browsing the internet. Somedays I think I am not 100% against of every website having a captcha...
评论 #43614244 未加载
评论 #43614237 未加载
knes大约 1 个月前
This is great. Especially debugging frontend issue on localhost or staging.<p>Also works flawlessly with augment code.com too!
picardo大约 1 个月前
I like this. It would be interesting to use it for when I need to use authenticated browser sessions.
lxe大约 1 个月前
This one also uses aria snapshots formatted as yaml. This will quickly exceed context limits.
plessas大约 1 个月前
thank you for this. Using my own browser helps me automate tasks on sites I &#x27;d typically get detected using automation. Works like a charm! Hope you continue to work on the repo.
jngiam1大约 1 个月前
Pretty cool, do you know of a version of this that supports the new remote MCP protocol
评论 #43614538 未加载
revskill大约 1 个月前
Can u expose the sdk as a react component to be used inside an app ?
mvdtnz大约 1 个月前
Is anyone successfully running MCPs &#x2F; Claude Desktop on Linux?
评论 #43618315 未加载
pknerd大约 1 个月前
So why do I need an editor(Cusror)? How does a non-coder use it?
评论 #43615496 未加载
xena大约 1 个月前
Do you respect robots.txt so administrators can block this tool?
评论 #43614885 未加载
评论 #43614852 未加载
cadence-大约 1 个月前
How does this compare to Anthropic&#x27;s Computer Use?
tuananh大约 1 个月前
i want to add this for my project (which use wasm) but rustlang&#x2F;socket2 WASI support is not merged yet. after that rust CDP will work.
评论 #43619051 未加载
jayunit大约 1 个月前
awesome! For the Cursor &#x2F; React &#x2F; Click to Add 2 example, can we also have it write a unit&#x2F;e2e regression test?
评论 #43616650 未加载
mrwww大约 1 个月前
How does it compare to playwright mcp?
graiz大约 1 个月前
works better than puppet mcp for me but having issues with keyboard events and actions on some websites.
johnpaulkiser大约 1 个月前
&gt; Private &gt; Since automation happens locally, your browser activity stays on your device and isn&#x27;t sent to remote servers.<p>I think this is bullshit. Isn&#x27;t the dom or whatever sent to the model api?
评论 #43613971 未加载
throwaway81523大约 1 个月前
Can these things automatically solve recaptcha? That&#x27;s the only AI browser feature that I have a real use for.
评论 #43616879 未加载
tntpreneur大约 1 个月前
Thanks but idea is ok but it is not working smoothly.
justanotheratom大约 1 个月前
neat, but instead of asking me to install browser extension, can you just bundle a browser in the MCP server?
tigrezno大约 1 个月前
this is the way
ndr大约 1 个月前
WARNING for Cursor users:<p>Cursor is currently stuck using an outdated snapshot of the VSCode Marketplace, meaning several extensions within Cursor remain affected by high-severity CVEs that have already been patched upstream in VSCode. As a result, Cursor users unknowingly remain vulnerable to known security issues. This issue has been acknowledged but remains unresolved: <a href="https:&#x2F;&#x2F;github.com&#x2F;getcursor&#x2F;cursor&#x2F;issues&#x2F;1602#issuecomment-2654870021">https:&#x2F;&#x2F;github.com&#x2F;getcursor&#x2F;cursor&#x2F;issues&#x2F;1602#issuecomment...</a><p>Given Cursor&#x27;s rising popularity, users should be aware of this gap in security updates. Until the Cursor team resolves the marketplace sync issue, caution is advised when using certain extensions.<p>I&#x27;ve flagged it here, apologies for the repost: <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=43609572">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=43609572</a>
评论 #43615018 未加载
评论 #43615720 未加载