[HN Gopher] Show HN: Web-eval-agent - Let the coding agent debug...
       ___________________________________________________________________
        
       Show HN: Web-eval-agent - Let the coding agent debug itself
        
       Hey HN! We've been building an MCP server to help AI-assisted web
       app developers by using browser agents to test whether changes made
       by an AI inside an editor actually work. We've been testing it on
       scenarios like verifying new flows in a UI, or checking that
       sending a chat request triggers a response. The idea is to let your
       coding agent both code and evaluate if what it did was correct.
       Here's a short demo with Cursor:
       https://www.youtube.com/watch?v=_AoQK-bwR0w  When building apps, we
       found the hardest part of AI-assisted coding isn't the coding--it's
       tedious point-and-click testing to see if things work. We got tired
       of this loop: open the app, click through flows, stare at the
       network tab, copy console errors to the editor, repeat. It felt
       obvious this should be AI-assisted too. If you can vibe-code, you
       should be able to vibe-test!  Some agents like Cline and Windsurf
       have browser integrations, but Cline's (via Anthropic Computer Use)
       felt slow and only reported console logs, and Windsurf's didn't
       work reliably yet. We got so tired of manually testing that we
       decided to fix it.  Our MCP server sits between your IDE agent
       (Cursor/Windsurf/Cline/Continue) and a Playwright-powered browser-
       use agent. It spins up the browser, navigates your app per
       instructions from the IDE agent, and sends back steps, console
       events, and network events so the IDE agent can assess the app's
       state.  We proxy Browser-use's original Claude calls and swap in
       Gemini Flash 2.0, cutting latency from ~8s - ~3s per step. We also
       cap console/network logs at 10,000 characters to stay within
       context limits, and filter out irrelevant logs (e.g., noisy XHR
       requests).  At the end, the browser agent outputs a summary like:
       Web Evaluation Report for http://localhost:5173        Task: delete
       an API key and evaluate UX       Steps: Home - Login - API Keys -
       Create Key - Delete Key       Flow tested successfully; UX had
       problems X, Y, Z...       Console (8)...   Network (13)...
       Timeline of events (57) ...       This gives the coding agent the
       ability to recognize the console and network errors, or any issues
       with clicking around, and have the coding agent fix them before
       returning back to the user. (There's a longer example in the README
       at https://github.com/Operative-Sh/web-eval-agent.)  Try it in
       Cursor / Cline / Windsurf / Claude Desktop: (macOS/Linux):
       curl -LSf https://operative.sh/install.sh -o install.sh       less
       -N install.sh   # inspect if you'd like       bash install.sh
       # installs uv + jq + Playwright + server       # then in
       Cursor/Cline/Windsurf/Continue: craft a prompt using the
       web_eval_agent tool       (For Windows, there's a 4-line manual
       install in the README.)  What we want to do next: pause/go for
       OAuth screens; save/load browser auth states; Playwright step
       recording for automated test creation and regression test creation;
       supporting Loveable / v0 / Bolt.new sites by offering a web
       version.  We'd love to hear your feedback, especially if you've
       experienced the pain of having to manually test changes happening
       in your web apps after making changes from inside your IDE, or if
       you've tried any alternative MCP tools for this that have worked
       well.  Try it out if you feel it'd be helpful for your workflow:
       https://github.com/Operative-Sh/web-eval-agent. (note: the server
       hits our operative.sh proxy to cover Gemini tokens. The MCP server
       itself is OSS; Anthropic base-URL support is coming soon. Free tier
       included; heavy users can grab the $10 plan to offset our model
       bill.)  Let us know what you think! Thanks for reading!
        
       Author : neversettles
       Score  : 56 points
       Date   : 2025-04-28 15:36 UTC (7 hours ago)
        
 (HTM) web link (github.com)
 (TXT) w3m dump (github.com)
        
       | GreenGames wrote:
       | This is very cool! Does your MCP server preserve
       | cookies/localStorage between steps, or would developers need to
       | manually script auth handshakes?
        
         | neversettles wrote:
         | Between steps it would preserve cookies, but atm when the
         | playwright browser launches, it starts with a fresh browser
         | state, so you'd have to o-auth to log in each time.
         | 
         | We're adding browser state persistence soon, hoping to enable
         | it so once you sign in with google once, it can stay signed in
         | on your local machine.
        
           | GreenGames wrote:
           | Oh okay thanks - that would be fire tbh
        
       | esafak wrote:
       | Is there a benchmark for this? If not, you ought to (crowd?)start
       | one for everybody's sake.
        
         | neversettles wrote:
         | We started with using browser-use because they had the best
         | evals: https://browser-use.com/posts/sota-technical-report
         | 
         | - but we found that Laminar came out with a better browser
         | agent (& a better eval): https://www.lmnr.ai/ so we're looking
         | to migrate over soon!
        
       | nico wrote:
       | Looks amazing. Congrats on the release
       | 
       | How does this compare to browser mcp (https://browsermcp.io/)?
        
         | neversettles wrote:
         | In browser MCP, looks like cursor controls each action along
         | the way, but actually what we wanted was a single browser agent
         | that had a high quality eval that could perform all the actions
         | independently (browser-use)
        
       | proc0 wrote:
       | Interesting. I see from the video example it took a lot of steps
       | and there is a lot of output for a simple task. I'm thinking this
       | probably doesn't scale very well and more complex tasks might
       | have performance challenges. I do think it's the right direction
       | for AI coding.
        
         | neversettles wrote:
         | Yeah, I suppose to esafak's point, perhaps a benchmark for
         | browser agent QA testing would be needed.
        
       ___________________________________________________________________
       (page generated 2025-04-28 23:00 UTC)