Trust & privacy
Security
usage.neural-forge.io is an independent, open-source mirror of public Cursor profile pages. Type a
@handle — we read the same stats anyone can see on cursor.com/@handle. Not affiliated with Cursor.Handle-only lookup
No login or accountsPublic profiles onlyOpen source (MIT)Cached snapshots
How lookup works
- You enter a public Cursor handle (for example
lauren). - The site fetches that handle's public profile page on cursor.com and parses the visible stats — tokens, activity, agents, models, streak, and join date.
- A snapshot is saved so the dashboard, Wrapped slides, and leaderboard load quickly. Snapshots refresh on a schedule (about every 12 hours) or when you hit refresh on a profile.
If a profile is private, missing, or not yet public on cursor.com, lookup fails — we do not bypass Cursor's visibility settings.
What gets stored
- Profile snapshots — JSON copies of public profile stats in Vercel Blob (or a local folder in dev). Used for
/{handle}pages and the leaderboard. - Billing stripped from public cache — spend, included usage, and on-demand billing fields are removed before a snapshot is written to shared storage. Public pages only show billing when the site operator has configured server-side access for their own handle.
- Leaderboard handle lists — handles discovered through searches plus a public analytics feed of active Cursor users. No emails or private account fields.
What we never do
- Ask visitors to paste a Cursor session token into the website
- Store Cursor credentials that visitors provide
- Create user accounts, passwords, or per-visitor sessions
- Invent or estimate usage numbers beyond the public profile
- Expose private billing or spend in the public profile cache
Community & sharing
- Community posts — optional updates builders submit at /community/post. Posts are moderated before they appear. Only the text, links, and tags you submit are stored.
- Share cards & Wrapped — OG images are generated from already-public profile stats. Download and X-share links point at your profile URL; they do not grant access to private data.
Caching, limits & removal
- Cached profiles auto-refresh about every 12 hours via a scheduled job, plus on-demand when you refresh a profile.
- API routes are rate limited (for example 20 reads and 10 refreshes per handle per minute) to keep the service stable.
- Because data is copied from a public cursor.com page, making your Cursor profile private or changing visibility on cursor.com is the authoritative way to stop new public reads. To request removal from this site's cache, open a GitHub issue.
Open source & reporting
The project is MIT-licensed on GitHub. You can audit how profiles are fetched in packages/core/ and how snapshots are cached in src/lib/profile-cache.ts.
Found a problem? Open a GitHub issue — please do not include session tokens, emails, or other private data.