Test a local app in Chrome and fix browser errors

A developer used Claude Code and Claude in Chrome to open a local app, exercise the UI, catch a routing bug, and repair it without manual browser work.

Done with Claude CodeClaude in Chrome· Intermediate

What they did
They connected the Claude in Chrome extension to Claude Code with `claude --chrome` or `/chrome`, which let the agent use their existing Chrome session and approved domains once. After asking it to add a dashboard table filter, the agent built the component, loaded the localhost page, tried the filter, and checked the browser for errors. When it saw that the URL query parameters were not changing, it adjusted the router logic and reloaded the page to verify the fix.
How it went
The filter example took three rounds to get right, and the developer never had to drive the browser themselves. The setup only works on the developer’s own machine with a visible Chrome window, and it is not meant for CI or headless runs. It also works in Chrome and Edge only.
Tips
Use it for local QA on your own workstation, not automated CI. Browser actions can burn through context quickly, so keep the task focused and ask the agent to pause before anything irreversible.

Try it yourself with Claude Code

I’m running a local app in Chrome. Start it with [start command], open [localhost URL], and walk through [feature or flow] as a real user would. Watch for console and network errors, trace any bug back to the code, and make the smallest fix needed. After each change, reload and retest until the flow works cleanly. Before you do anything irreversible, such as submitting a form, deleting data, or making a payment, stop and ask me to approve it. Then summarize what you changed and what you checked.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on test-lab.ai ↗

3 more reports of this

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.