The build in Cursor browser is not working properly. It does not render everything correctly. My try to use Google Chrome failed, because the app tells me it can’t find Google Chrome on my machine, even though it is updated and there for years. Currently opening the app preview in Google Chrome and not in Cursor directly, which is suuuuupper frustrating. I need the Cursor browser back or I’m gone.
Hey, thanks for the report. Just a couple of things to clarify:
What exactly isn’t rendering - backgrounds, animations, specific CSS effects, or is it everything? A screenshot would really help: how it looks in Cursor vs in Chrome.
Regarding “Chrome not found” - can you check in Cursor settings what path to the browser is specified? Cursor Settings > Tools & MCP > Browser, there should be a path to Chrome.
@brales - are you also having rendering issues or is it just that Chrome isn’t found?
A similar bug with backgrounds was fixed in 2.4, so I want to understand if this is the same issue returning or something new.
1. I set Chrome in Cursor Settings > Tools & MCP > Browser. When I ask in chat to explore a web page, the AI responds like this: “I can’t open that page directly from here. If you open it, I can guide you through the steps or analyze a screenshot or HTML that you paste.” So it can’t access the browser at all. Chrome is installed on my computer.
2. If I set “Browser Tab” in Cursor Settings > Tools & MCP > Browser, and ask it to open the browser in a tab, it responds like this: “Sorry, I don’t have access to a shared browser from here.” If I open the tab with the browser and navigate to a web page and ask it to go to the developer tools console, it responds, “I can’t see your browser’s developer tools console.”
Basically, I can’t get the AI to access the browser in any way to help me with code development.
For me the app is missing randomly buttons, while opening localhost in Google Chrome (externally) works perfectly. Tried to change many things like changing the browser or HTTPS, nothing worked. For me is that e.g. one button is not rendered in the browser in Cursor, but in Chrome externally.
Thanks, could you send screenshots of how the page looks in the Cursor browser vs in Chrome? We need to see exactly which elements are missing.
Also, check in Cursor Settings > Tools & MCP > Browser what browser path is set there. Try switching between “Browser Tab” and “Chrome”. Does anything change?
@brales, it looks like you might have a separate issue with browser automation, not rendering. When you set Chrome in settings and ask the AI to open a page, does it try to open a Chrome window, or does it run in the Browser Tab? Can you share a screenshot of what you see in the browser tab when you open localhost?
Hi Dean, here are my answers to your questions: When you set Chrome in settings and ask the AI to open a page, does it try to open a Chrome window, or does it run in the Browser Tab?
It doesn’t open Chrome or the Browser Tab. If I ask it to open a page (e.g., google.com), it responds like this:
—-
I don’t have access to tools that open a visual browser window you can see. I can fetch and display web content, but not open a browser.
Here’s the content from Google.com that I retrieved:
Google Homepage Content:
The page shows:
Google logo at the top
Navigation bar with links to:
Gmail
Images (Grafika)
Google Store
About Google
Main search area (the central search box)
Language: The page is in Polish (Poland)
Cookie consent dialog asking to accept/reject cookies
Footer with links to:
Privacy (Prywatność)
Terms (Warunki)
Settings (Ustawienia)
Advanced Search (Szukanie zaawansowane)
Help (Pomoc)
Cookie Dialog Content:
The page displays a cookie consent dialog in Polish explaining:
How Google uses cookies and data
Options to “Accept all”, “Reject all”, or “More options”
Links to privacy policy and terms
The page appears localized for Poland, which is why the text is in Polish.
I can fetch content from any URL and show it here, but I cannot open a visual browser window. If you need to interact with a page (click buttons, fill forms), I’d need browser automation tools, which aren’t available here.
Would you like me to fetch content from a different URL or format this differently?
—-
When the AI responded correctly (until about three updates ago, if I remember correctly), it was able to see the content on both Chrome and the Browser Tab and could also analyze the browser’s developer tools to analyze the open pages. This behavior is no longer available.
I have attached a screenshot of what I see when I go to localhost (I have activated Wampserver as the local server).
