{"slug": "chrome-devtools-mcp", "name": "Chrome DevTools MCP", "category": "browsers", "category_name": "Browsers and extensions", "description": "Official MCP server that lets AI coding agents control and inspect Chrome via DevTools.", "recommenders": 7, "commenters": 11, "warners": 2, "score": 4.756, "rising": -2.29, "first": "2025-11-06", "latest": "2026-08-28", "why": "Liked for letting coding agents drive Chrome to check their own work, read network requests as context, and replace slower web-driver or Playwright setups.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 7, "neutral": 2, "compare": 0, "warn": 2}, "spark": [1, 0, 0, 2, 2, 1, 0, 0, 1, 0, 0, 0], "listed": true, "loved": ["Agent self-testing", "Network inspection"], "caveat": "Token heavy", "short_quote": "would highly recommend using the Chrome Devtools MCP to be able to attach the network requests", "icon": true, "domain": null, "github_repo": "github.com/chromedevtools/chrome-devtools-mcp", "official_url": "https://github.com/ChromeDevTools/chrome-devtools-mcp", "pricing": "open source", "aliases": ["Chrome DevTools MCP", "chrome-devtools MCP"], "verdict": [{"text": "Liked for letting coding agents drive Chrome to check their own work", "cites": [{"id": 733503636, "item_id": 45843977, "url": "https://news.ycombinator.com/item?id=45843977", "by": "timinou", "date": "2025-11-06", "stance": "recommend", "quote": "the LLM can inspect the behaviour using the chrome dev tools MCP and check its work much more effectively", "excerpt": "I don't buy it either. I've been building my own backend framework for the past 2.5 years, and even though it's a DSL over Python and there's no documentation online and barely one in my computer, Claude Code understands it with enough usage examples in my codebase. In front-end as well\u2014I've been able to go much farther for simple projects using alpine than more complex frameworks. For big products I use Elm, which i", "context": "", "use_case": "LLM inspecting front-end behavior", "switched_from": "", "story": "Dead Framework Theory"}, {"id": 753187330, "item_id": 47074208, "url": "https://news.ycombinator.com/item?id=47074208", "by": "ahmed_sulajman", "date": "2026-02-19", "stance": "recommend", "quote": "So that it can take screenshots and compare to expected design", "excerpt": "Not as much for the UX, but at least for UI when you need to implement designs, I use Claude Code with Figma MCP and Chrome Dev tools MCP. So that it can take screenshots and compare to expected design as part of the acceptance criteria. For a more targeted fine tuning of the UI I also started using Agentation https://github.com/benjitaylor/agentation if I'm working on React-based app", "context": "", "use_case": "screenshots for UI checking", "switched_from": "", "story": "Ask HN: How do you employ LLMs for UI development?"}, {"id": 776739700, "item_id": 48546231, "url": "https://news.ycombinator.com/item?id=48546231", "by": "garethsprice", "date": "2026-06-15", "stance": "neutral", "quote": "overnight I use the Chrome devtools MCP  and have the above setup fuzz-test as a user", "excerpt": "Using OpenCode + OhMyOpenCode + Qwen 3.6 35B-A3B Q_4_KM on an Ada 4000 (20GB VRAM) at 55 tok/sec for generation (slower than it sounds as OpenCode has a bunch of context it adds). Meaning to check out pi when I get a minute as I hear that one mentioned a lot lately. I am using Opus to generate plans that the local agent then follows, then validated by Opus. So I'm not at 100% local but these models are increasingly p", "context": "", "use_case": "overnight automated fuzz-testing", "switched_from": "", "story": "Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?"}, {"id": 758603794, "item_id": 47412737, "url": "https://news.ycombinator.com/item?id=47412737", "by": "s1mon", "date": "2026-03-17", "stance": "neutral", "quote": "tooling through Google Chrome Devtools MCP and Playwright to extract things and control the browser", "excerpt": "I've been using Codex (GPT-5.4 extra high) to code custom FeatureScript in Onshape (3D mechanical CAD software). It's challenging to get it to do TDD that involves any visual reasoning. At the moment I've got tooling through Google Chrome Devtools MCP and Playwright to extract things and control the browser and I use some custom features which help with formatting and controlling debugging outputs (text and visual ov", "context": "", "use_case": "browser control for agent", "switched_from": "", "story": "Claude Tips for 3D Work"}]}, {"text": ", read network requests as context", "cites": [{"id": 766906960, "item_id": 47931685, "url": "https://news.ycombinator.com/item?id=47931685", "by": "safehuss", "date": "2026-04-28", "stance": "recommend", "quote": "would highly recommend using the Chrome Devtools MCP to be able to attach the network requests", "excerpt": "This is awesome! Worked on something similar albeit a different industry. For the more challenging scrapes, would highly recommend using the Chrome Devtools MCP to be able to attach the network requests, being made by the browser to the site, as context for your agent/LLM chat - this approach really helped me to write a solid API-based scraper (also using curl_cffi) and bypassed the old tedious playwright-based appro", "context": "", "use_case": "web scraping with LLM context", "switched_from": "Playwright", "story": "Scraping 241 UK council planning portals \u2013 2.6M decisions so far"}]}, {"text": ", and replace slower web-driver or Playwright setups.", "cites": [{"id": 780353216, "item_id": 48772076, "url": "https://news.ycombinator.com/item?id=48772076", "by": "bel8", "date": "2026-07-03", "stance": "recommend", "quote": "I have been using Chrome's official MCP devtools server since Nov 2025.", "excerpt": "I have been using Chrome's official MCP devtools server since Nov 2025. https://github.com/ChromeDevTools/chrome-devtools-mcp Before that I used Chrome web drivers but MCP is faster and more capable. I also instruct LLMs to test my pages on Firefox using its official MCP to make sure they work in Firefox too: https://github.com/mozilla/firefox-devtools-mcp Now I will add Safari to the compatibility tests. cool", "context": "", "use_case": "LLM page testing in Chrome", "switched_from": "Chrome web drivers", "story": "The Safari MCP server for web developers"}, {"id": 766906960, "item_id": 47931685, "url": "https://news.ycombinator.com/item?id=47931685", "by": "safehuss", "date": "2026-04-28", "stance": "recommend", "quote": "would highly recommend using the Chrome Devtools MCP to be able to attach the network requests", "excerpt": "This is awesome! Worked on something similar albeit a different industry. For the more challenging scrapes, would highly recommend using the Chrome Devtools MCP to be able to attach the network requests, being made by the browser to the site, as context for your agent/LLM chat - this approach really helped me to write a solid API-based scraper (also using curl_cffi) and bypassed the old tedious playwright-based appro", "context": "", "use_case": "web scraping with LLM context", "switched_from": "Playwright", "story": "Scraping 241 UK council planning portals \u2013 2.6M decisions so far"}]}, {"text": " The main caveats are heavy token use for some users", "cites": [{"id": 758270816, "item_id": 47391926, "url": "https://news.ycombinator.com/item?id=47391926", "by": "rossvc", "date": "2026-03-15", "stance": "warn", "quote": "I've been using the DevTools MCP for months now, but it's extremely token heavy.", "excerpt": "I've been using the DevTools MCP for months now, but it's extremely token heavy. Is there an alternative that provides the same amount of detail when it comes to reading back network requests?", "context": "", "use_case": "reading network requests", "switched_from": "", "story": "Let your Coding Agent debug the browser session with Chrome DevTools MCP"}]}, {"text": " and a broad security surface that is only as safe as the browser sandbox.", "cites": [{"id": 791638609, "item_id": 49477413, "url": "https://news.ycombinator.com/item?id=49477413", "by": "nameless912", "date": "2026-08-28", "stance": "warn", "quote": "the Chrome Dev tools MCP is less discriminating and is only as secure as a browser sandbox with full JS access", "excerpt": "This article smells AI generated, which is _very_ funny given the argument being made. Either way, while this is true in the absolute, this is the value of building good MCP servers: they should expose only exactly the surface you expect your agent to need, and adding functionality should be carefully considered. The best MCP servers I use day to day (Cloudflare sticks out) do a really good job of exposing only what ", "context": "", "use_case": "", "switched_from": "", "story": "AI Agent Has Root"}]}], "praise": [{"label": "Agent self-testing", "summary": "Lets LLM agents inspect pages, take screenshots and verify UI or behavior as they code.", "quotes": [{"id": 733503636, "item_id": 45843977, "url": "https://news.ycombinator.com/item?id=45843977", "by": "timinou", "date": "2025-11-06", "stance": "recommend", "quote": "the LLM can inspect the behaviour using the chrome dev tools MCP and check its work much more effectively", "excerpt": "I don't buy it either. I've been building my own backend framework for the past 2.5 years, and even though it's a DSL over Python and there's no documentation online and barely one in my computer, Claude Code understands it with enough usage examples in my codebase. In front-end as well\u2014I've been able to go much farther for simple projects using alpine than more complex frameworks. For big products I use Elm, which i", "context": "", "use_case": "LLM inspecting front-end behavior", "switched_from": "", "story": "Dead Framework Theory"}, {"id": 753187330, "item_id": 47074208, "url": "https://news.ycombinator.com/item?id=47074208", "by": "ahmed_sulajman", "date": "2026-02-19", "stance": "recommend", "quote": "So that it can take screenshots and compare to expected design", "excerpt": "Not as much for the UX, but at least for UI when you need to implement designs, I use Claude Code with Figma MCP and Chrome Dev tools MCP. So that it can take screenshots and compare to expected design as part of the acceptance criteria. For a more targeted fine tuning of the UI I also started using Agentation https://github.com/benjitaylor/agentation if I'm working on React-based app", "context": "", "use_case": "screenshots for UI checking", "switched_from": "", "story": "Ask HN: How do you employ LLMs for UI development?"}, {"id": 776739700, "item_id": 48546231, "url": "https://news.ycombinator.com/item?id=48546231", "by": "garethsprice", "date": "2026-06-15", "stance": "neutral", "quote": "overnight I use the Chrome devtools MCP  and have the above setup fuzz-test as a user", "excerpt": "Using OpenCode + OhMyOpenCode + Qwen 3.6 35B-A3B Q_4_KM on an Ada 4000 (20GB VRAM) at 55 tok/sec for generation (slower than it sounds as OpenCode has a bunch of context it adds). Meaning to check out pi when I get a minute as I hear that one mentioned a lot lately. I am using Opus to generate plans that the local agent then follows, then validated by Opus. So I'm not at 100% local but these models are increasingly p", "context": "", "use_case": "overnight automated fuzz-testing", "switched_from": "", "story": "Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?"}]}, {"label": "Network inspection", "summary": "Network requests can be passed to the agent as context, which helped one user build API-based scrapers.", "quotes": [{"id": 766906960, "item_id": 47931685, "url": "https://news.ycombinator.com/item?id=47931685", "by": "safehuss", "date": "2026-04-28", "stance": "recommend", "quote": "would highly recommend using the Chrome Devtools MCP to be able to attach the network requests", "excerpt": "This is awesome! Worked on something similar albeit a different industry. For the more challenging scrapes, would highly recommend using the Chrome Devtools MCP to be able to attach the network requests, being made by the browser to the site, as context for your agent/LLM chat - this approach really helped me to write a solid API-based scraper (also using curl_cffi) and bypassed the old tedious playwright-based appro", "context": "", "use_case": "web scraping with LLM context", "switched_from": "Playwright", "story": "Scraping 241 UK council planning portals \u2013 2.6M decisions so far"}, {"id": 758270816, "item_id": 47391926, "url": "https://news.ycombinator.com/item?id=47391926", "by": "rossvc", "date": "2026-03-15", "stance": "warn", "quote": "I've been using the DevTools MCP for months now, but it's extremely token heavy.", "excerpt": "I've been using the DevTools MCP for months now, but it's extremely token heavy. Is there an alternative that provides the same amount of detail when it comes to reading back network requests?", "context": "", "use_case": "reading network requests", "switched_from": "", "story": "Let your Coding Agent debug the browser session with Chrome DevTools MCP"}]}, {"label": "Faster, more capable", "summary": "Described as faster and more capable than Chrome web drivers and more reliable than other DevTools MCPs.", "quotes": [{"id": 780353216, "item_id": 48772076, "url": "https://news.ycombinator.com/item?id=48772076", "by": "bel8", "date": "2026-07-03", "stance": "recommend", "quote": "I have been using Chrome's official MCP devtools server since Nov 2025.", "excerpt": "I have been using Chrome's official MCP devtools server since Nov 2025. https://github.com/ChromeDevTools/chrome-devtools-mcp Before that I used Chrome web drivers but MCP is faster and more capable. I also instruct LLMs to test my pages on Firefox using its official MCP to make sure they work in Firefox too: https://github.com/mozilla/firefox-devtools-mcp Now I will add Safari to the compatibility tests. cool", "context": "", "use_case": "LLM page testing in Chrome", "switched_from": "Chrome web drivers", "story": "The Safari MCP server for web developers"}, {"id": 758263824, "item_id": 47391489, "url": "https://news.ycombinator.com/item?id=47391489", "by": "boomskats", "date": "2026-03-15", "stance": "recommend", "quote": "It's more reliable and token efficient than other devtools protocol MCPs i've tried.", "excerpt": "Been using this one for a while, mostly with codex on opencode. It's more reliable and token efficient than other devtools protocol MCPs i've tried. Favourite unexpected use case for me was telling gemini to use it as a SVG editing repl, where it was able to produce some fantastic looking custom icons for me after 3-4 generate/refresh/screenshot iterations. Also works very nicely with electron apps, both reverse engi", "context": "", "use_case": "agent browser debugging", "switched_from": "", "story": "Let your Coding Agent debug the browser session with Chrome DevTools MCP"}]}, {"label": "Browser control", "summary": "Used alongside Playwright to extract data and control the browser in custom workflows.", "quotes": [{"id": 758603794, "item_id": 47412737, "url": "https://news.ycombinator.com/item?id=47412737", "by": "s1mon", "date": "2026-03-17", "stance": "neutral", "quote": "tooling through Google Chrome Devtools MCP and Playwright to extract things and control the browser", "excerpt": "I've been using Codex (GPT-5.4 extra high) to code custom FeatureScript in Onshape (3D mechanical CAD software). It's challenging to get it to do TDD that involves any visual reasoning. At the moment I've got tooling through Google Chrome Devtools MCP and Playwright to extract things and control the browser and I use some custom features which help with formatting and controlling debugging outputs (text and visual ov", "context": "", "use_case": "browser control for agent", "switched_from": "", "story": "Claude Tips for 3D Work"}, {"id": 752048848, "item_id": 47003053, "url": "https://news.ycombinator.com/item?id=47003053", "by": "Kim_Bruning", "date": "2026-02-13", "stance": "recommend", "quote": "I've been using chrome-devtools fwiw", "excerpt": "I've seen claude --chrome , but there's also at least 2 mcps that can puppeteer chrome. I've been using chrome-devtools fwiw", "context": "", "use_case": "letting AI agents test web apps", "switched_from": "", "story": "Ask HN: Why can't Codex/Claude compile app and test that changes worked?"}]}, {"label": "Electron support", "summary": "One user says it works well with Electron apps for reverse engineering and extending them.", "quotes": [{"id": 758263824, "item_id": 47391489, "url": "https://news.ycombinator.com/item?id=47391489", "by": "boomskats", "date": "2026-03-15", "stance": "recommend", "quote": "It's more reliable and token efficient than other devtools protocol MCPs i've tried.", "excerpt": "Been using this one for a while, mostly with codex on opencode. It's more reliable and token efficient than other devtools protocol MCPs i've tried. Favourite unexpected use case for me was telling gemini to use it as a SVG editing repl, where it was able to produce some fantastic looking custom icons for me after 3-4 generate/refresh/screenshot iterations. Also works very nicely with electron apps, both reverse engi", "context": "", "use_case": "agent browser debugging", "switched_from": "", "story": "Let your Coding Agent debug the browser session with Chrome DevTools MCP"}]}], "caveats": [{"label": "Token heavy", "summary": "One long-time user calls it extremely token heavy, though another found it more token efficient than alternatives.", "quotes": [{"id": 758270816, "item_id": 47391926, "url": "https://news.ycombinator.com/item?id=47391926", "by": "rossvc", "date": "2026-03-15", "stance": "warn", "quote": "I've been using the DevTools MCP for months now, but it's extremely token heavy.", "excerpt": "I've been using the DevTools MCP for months now, but it's extremely token heavy. Is there an alternative that provides the same amount of detail when it comes to reading back network requests?", "context": "", "use_case": "reading network requests", "switched_from": "", "story": "Let your Coding Agent debug the browser session with Chrome DevTools MCP"}, {"id": 758263824, "item_id": 47391489, "url": "https://news.ycombinator.com/item?id=47391489", "by": "boomskats", "date": "2026-03-15", "stance": "recommend", "quote": "It's more reliable and token efficient than other devtools protocol MCPs i've tried.", "excerpt": "Been using this one for a while, mostly with codex on opencode. It's more reliable and token efficient than other devtools protocol MCPs i've tried. Favourite unexpected use case for me was telling gemini to use it as a SVG editing repl, where it was able to produce some fantastic looking custom icons for me after 3-4 generate/refresh/screenshot iterations. Also works very nicely with electron apps, both reverse engi", "context": "", "use_case": "agent browser debugging", "switched_from": "", "story": "Let your Coding Agent debug the browser session with Chrome DevTools MCP"}]}, {"label": "Broad security surface", "summary": "Exposes full JS access rather than a narrow tool set, so it is only as secure as the browser sandbox.", "quotes": [{"id": 791638609, "item_id": 49477413, "url": "https://news.ycombinator.com/item?id=49477413", "by": "nameless912", "date": "2026-08-28", "stance": "warn", "quote": "the Chrome Dev tools MCP is less discriminating and is only as secure as a browser sandbox with full JS access", "excerpt": "This article smells AI generated, which is _very_ funny given the argument being made. Either way, while this is true in the absolute, this is the value of building good MCP servers: they should expose only exactly the surface you expect your agent to need, and adding functionality should be carefully considered. The best MCP servers I use day to day (Cloudflare sticks out) do a really good job of exposing only what ", "context": "", "use_case": "", "switched_from": "", "story": "AI Agent Has Root"}]}], "who": [], "tags": ["for AI coding agents", "for UI design checks", "for web scraping", "for Electron apps"], "use_cases": ["llm page testing in chrome", "overnight automated fuzz-testing", "web scraping with llm context", "agent browser control", "browser control for agent", "reading network requests"], "compared": [{"slug": "playwright", "name": "Playwright", "co": 3, "switched_from": 1, "switched_to": 0}, {"slug": "claude-code", "name": "Claude Code", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "codex", "name": "Codex", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "opencode", "name": "OpenCode", "co": 2, "switched_from": 0, "switched_to": 0}], "makers": [{"id": 758273632, "item_id": 47392102, "url": "https://news.ycombinator.com/item?id=47392102", "by": "paulirish", "date": "2026-03-15", "stance": "recommend", "quote": "The DevTools MCP project just recently landed a standalone CLI", "excerpt": "The DevTools MCP project just recently landed a standalone CLI: https://github.com/ChromeDevTools/chrome-devtools-mcp/blob/main/skills/chrome-devtools-cli/SKILL.md Great news to all of us keenly aware of MCP's wild token costs. ;) The CLI hasn't been announced yet (sorry guys!), but it is shipping in the latest v0.20.0 release. (Disclaimer: I used to work on the DevTools team. And I still do, too)", "context": "used to work on the DevTools team", "use_case": "agent browser debugging", "switched_from": "", "story": "Let your Coding Agent debug the browser session with Chrome DevTools MCP"}], "threads": [{"url": "https://news.ycombinator.com/item?id=47390817", "title": "Let your Coding Agent debug the browser session with Chrome DevTools MCP", "mentions": 3}, {"url": "https://news.ycombinator.com/item?id=49477311", "title": "AI Agent Has Root", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48769639", "title": "The Safari MCP server for web developers", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48542100", "title": "Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47931470", "title": "Scraping 241 UK council planning portals \u2013 2.6M decisions so far", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47503112", "title": "What are the best headless browsers?", "mentions": 1}], "sentiment_by_month": {"2025-11": {"recommend": 1}, "2025-12": {}, "2026-01": {}, "2026-02": {"recommend": 2}, "2026-03": {"recommend": 2, "warn": 1, "neutral": 1}, "2026-04": {"recommend": 1}, "2026-05": {}, "2026-06": {"neutral": 1}, "2026-07": {"recommend": 1}, "2026-08": {"warn": 1}, "2026-09": {}, "2026-10": {}}, "quotes": [{"id": 791638609, "item_id": 49477413, "url": "https://news.ycombinator.com/item?id=49477413", "by": "nameless912", "date": "2026-08-28", "stance": "warn", "quote": "the Chrome Dev tools MCP is less discriminating and is only as secure as a browser sandbox with full JS access", "excerpt": "This article smells AI generated, which is _very_ funny given the argument being made. Either way, while this is true in the absolute, this is the value of building good MCP servers: they should expose only exactly the surface you expect your agent to need, and adding functionality should be carefully considered. The best MCP servers I use day to day (Cloudflare sticks out) do a really good job of exposing only what ", "context": "", "use_case": "", "switched_from": "", "story": "AI Agent Has Root"}, {"id": 780353216, "item_id": 48772076, "url": "https://news.ycombinator.com/item?id=48772076", "by": "bel8", "date": "2026-07-03", "stance": "recommend", "quote": "I have been using Chrome's official MCP devtools server since Nov 2025.", "excerpt": "I have been using Chrome's official MCP devtools server since Nov 2025. https://github.com/ChromeDevTools/chrome-devtools-mcp Before that I used Chrome web drivers but MCP is faster and more capable. I also instruct LLMs to test my pages on Firefox using its official MCP to make sure they work in Firefox too: https://github.com/mozilla/firefox-devtools-mcp Now I will add Safari to the compatibility tests. cool", "context": "", "use_case": "LLM page testing in Chrome", "switched_from": "Chrome web drivers", "story": "The Safari MCP server for web developers"}, {"id": 776739700, "item_id": 48546231, "url": "https://news.ycombinator.com/item?id=48546231", "by": "garethsprice", "date": "2026-06-15", "stance": "neutral", "quote": "overnight I use the Chrome devtools MCP  and have the above setup fuzz-test as a user", "excerpt": "Using OpenCode + OhMyOpenCode + Qwen 3.6 35B-A3B Q_4_KM on an Ada 4000 (20GB VRAM) at 55 tok/sec for generation (slower than it sounds as OpenCode has a bunch of context it adds). Meaning to check out pi when I get a minute as I hear that one mentioned a lot lately. I am using Opus to generate plans that the local agent then follows, then validated by Opus. So I'm not at 100% local but these models are increasingly p", "context": "", "use_case": "overnight automated fuzz-testing", "switched_from": "", "story": "Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?"}, {"id": 766906960, "item_id": 47931685, "url": "https://news.ycombinator.com/item?id=47931685", "by": "safehuss", "date": "2026-04-28", "stance": "recommend", "quote": "would highly recommend using the Chrome Devtools MCP to be able to attach the network requests", "excerpt": "This is awesome! Worked on something similar albeit a different industry. For the more challenging scrapes, would highly recommend using the Chrome Devtools MCP to be able to attach the network requests, being made by the browser to the site, as context for your agent/LLM chat - this approach really helped me to write a solid API-based scraper (also using curl_cffi) and bypassed the old tedious playwright-based appro", "context": "", "use_case": "web scraping with LLM context", "switched_from": "Playwright", "story": "Scraping 241 UK council planning portals \u2013 2.6M decisions so far"}, {"id": 760055506, "item_id": 47503469, "url": "https://news.ycombinator.com/item?id=47503469", "by": "hwartig", "date": "2026-03-24", "stance": "recommend", "quote": "you also should take a look at https://developer.chrome.com/blog/chrome-devtools-mcp", "excerpt": "I'd recommend chrome(ium) headless https://developer.chrome.com/docs/chromium/headless . It works with various drivers like puppeteer, selenium and playwright for some added convenience. At https://checklyhq.com we and our customers use it heavily and reliably through playwright for various monitoring, testing and automation use cases. As you want to use it from an Agent you also should take a look at https://develop", "context": "", "use_case": "agent browser control", "switched_from": "", "story": "What are the best headless browsers?"}, {"id": 758603794, "item_id": 47412737, "url": "https://news.ycombinator.com/item?id=47412737", "by": "s1mon", "date": "2026-03-17", "stance": "neutral", "quote": "tooling through Google Chrome Devtools MCP and Playwright to extract things and control the browser", "excerpt": "I've been using Codex (GPT-5.4 extra high) to code custom FeatureScript in Onshape (3D mechanical CAD software). It's challenging to get it to do TDD that involves any visual reasoning. At the moment I've got tooling through Google Chrome Devtools MCP and Playwright to extract things and control the browser and I use some custom features which help with formatting and controlling debugging outputs (text and visual ov", "context": "", "use_case": "browser control for agent", "switched_from": "", "story": "Claude Tips for 3D Work"}, {"id": 758270816, "item_id": 47391926, "url": "https://news.ycombinator.com/item?id=47391926", "by": "rossvc", "date": "2026-03-15", "stance": "warn", "quote": "I've been using the DevTools MCP for months now, but it's extremely token heavy.", "excerpt": "I've been using the DevTools MCP for months now, but it's extremely token heavy. Is there an alternative that provides the same amount of detail when it comes to reading back network requests?", "context": "", "use_case": "reading network requests", "switched_from": "", "story": "Let your Coding Agent debug the browser session with Chrome DevTools MCP"}, {"id": 758263824, "item_id": 47391489, "url": "https://news.ycombinator.com/item?id=47391489", "by": "boomskats", "date": "2026-03-15", "stance": "recommend", "quote": "It's more reliable and token efficient than other devtools protocol MCPs i've tried.", "excerpt": "Been using this one for a while, mostly with codex on opencode. It's more reliable and token efficient than other devtools protocol MCPs i've tried. Favourite unexpected use case for me was telling gemini to use it as a SVG editing repl, where it was able to produce some fantastic looking custom icons for me after 3-4 generate/refresh/screenshot iterations. Also works very nicely with electron apps, both reverse engi", "context": "", "use_case": "agent browser debugging", "switched_from": "", "story": "Let your Coding Agent debug the browser session with Chrome DevTools MCP"}, {"id": 753187330, "item_id": 47074208, "url": "https://news.ycombinator.com/item?id=47074208", "by": "ahmed_sulajman", "date": "2026-02-19", "stance": "recommend", "quote": "So that it can take screenshots and compare to expected design", "excerpt": "Not as much for the UX, but at least for UI when you need to implement designs, I use Claude Code with Figma MCP and Chrome Dev tools MCP. So that it can take screenshots and compare to expected design as part of the acceptance criteria. For a more targeted fine tuning of the UI I also started using Agentation https://github.com/benjitaylor/agentation if I'm working on React-based app", "context": "", "use_case": "screenshots for UI checking", "switched_from": "", "story": "Ask HN: How do you employ LLMs for UI development?"}, {"id": 752048848, "item_id": 47003053, "url": "https://news.ycombinator.com/item?id=47003053", "by": "Kim_Bruning", "date": "2026-02-13", "stance": "recommend", "quote": "I've been using chrome-devtools fwiw", "excerpt": "I've seen claude --chrome , but there's also at least 2 mcps that can puppeteer chrome. I've been using chrome-devtools fwiw", "context": "", "use_case": "letting AI agents test web apps", "switched_from": "", "story": "Ask HN: Why can't Codex/Claude compile app and test that changes worked?"}, {"id": 733503636, "item_id": 45843977, "url": "https://news.ycombinator.com/item?id=45843977", "by": "timinou", "date": "2025-11-06", "stance": "recommend", "quote": "the LLM can inspect the behaviour using the chrome dev tools MCP and check its work much more effectively", "excerpt": "I don't buy it either. I've been building my own backend framework for the past 2.5 years, and even though it's a DSL over Python and there's no documentation online and barely one in my computer, Claude Code understands it with enough usage examples in my codebase. In front-end as well\u2014I've been able to go much farther for simple projects using alpine than more complex frameworks. For big products I use Elm, which i", "context": "", "use_case": "LLM inspecting front-end behavior", "switched_from": "", "story": "Dead Framework Theory"}]}