Where does the bug appear (feature/product)?
Somewhere else…
Describe the Bug
the docs domain migration broke cursor’s llms.txt for anyone holding the old URL. https://docs.cursor.com/llms.txt redirects to https://cursor.com/docs, which is an HTML page. until the docs moved, that was the correct llms.txt location, so any AI tool or agent that saved it now ingests a React app shell instead of markdown. https://cursor.com/docs/llms.txt also returns the HTML shell. the real file lives at https://cursor.com/llms.txt and works fine.
Steps to Reproduce
$ curl -sL https://docs.cursor.com/llms.txt | head -c 15
<!DOCTYPE html>
$ curl -sL https://cursor.com/llms.txt | head -c 15
# Cursor
or open both URLs in a browser: the first renders a webpage, the second returns plain markdown. also reproducible with npx llms-txt-check https://docs.cursor.com, a checker i built for exactly this class of breakage.
Expected Behavior
docs.cursor.com/llms.txt (and ideally cursor.com/docs/llms.txt) should redirect to https://cursor.com/llms.txt so existing consumers of the old URL get the markdown file instead of HTML. same for llms-full.txt if the old domain served one.
Operating System
Windows 10/11
MacOS
Linux
Other
Version Information
Other (website bug
For AI issues: which model did you use?
this is about cursor.com / docs.cursor.com routing
For AI issues: add Request ID with privacy disabled
n/a
Additional Information
for context, i maintain docusaurus-plugin-copy-page-button (runs on 28 docs sites; react native and pnpm are the big ones, puppeteer too) so i spend a lot of time in this corner of docs tooling. running npx [email protected] https://cursor.com as a CI step after docs deploys would catch this kind of breakage before agents hit it, it exits nonzero when the file or any listed url stops serving.
Does this stop you from using Cursor
No - Cursor works, but with this issue