I’m not sure if this is a silly suggestion - but when debugging detailed front end issues, I find myself reverting to extracting JS console logs (copy and paste) to help cursor understand more about what’s going on in the browser.
I understand various tools are around to help extracting Js console logs but most seems pretty complex to setup and configure. Has anyone built something simple to allow JS console logs to be extracted or fed to cursor? (dong it manually is pretty tedious particularly with apparent limitation of dev console in not auto expanding console logs or allowing easy export of full logs including JSON message structure) - hopefully I’m just overlooking some existing feature?
Hi Paul, I found that the recent introduction of ‘Debug’ mode (switch from agent in the cursor interface) when used in combination with the ‘@Browser’ directive to load my app inside cursor itself (I zoomed the browser view out to ~70% so I can see enough / use is on my screen) - gave cursor the ability to look deeply enough inside the app and sniff out additional debugging info through its own internal tooling approach, to satisfy my needs. I use Remote SSH to connect to my server but that was not a blocker. So perhaps give that a go. Switch to debug mode. Use the cursor browser to load up the page or area of your app you are struggling with. Give Cursor’s debug agent a clear description of the problem and let it do its thing. That has been working well for me recently.
even using `@browser` in debug mode, it still can’t access my console logs, that’s very weird! I think I’ll just use an MCP for now. Thanks for your help though!
Yes I asked cursor itself that question - can’t recall the specifics (it was a few weeks ago) but I gave it something along the lines of a prompt “If using the local @Browser with debug mode tooling, are you able to see the console logs in the local cursor browser?“ and I recall it giving me a positive response that it was able to see those - but you may want to quiz cursor further about this and ask it how it does it. Logically (in my mind at least), if it controls a local browser inside the cursor window itself it would seem to make sense that it could see (its own browser) logs. I know that before finding this approach I was researching tooling to scan JS console logs in Chrome and there were various third party platforms that offered that but it felt super complex and messy (signing up for third party tools - adding sofware to server - ‘connecting’ that software to the app etc). For me, however it achieves it - its seems to be pretty effective.