This page is a plain-English description of how githusb handles data. It’s not a legal document and doesn’t claim to be. If you need one before using the service, please don’t use the service.
What we read
- Public data from api.github.com about the GitHub handle you enter: profile, public events, public repositories. We do not request or access any private data.
- Your browser’s standard request metadata (IP, user agent, referrer) as handled by our hosting provider (Vercel) for abuse prevention.
What we send to third parties
- A structured summary of the GitHub signals we extracted is sent to a language-model provider to produce the written analysis. Today: Groq (
llama-3.3-70b-versatile). The provider may process and briefly log this request per its own terms. - No cookies, no analytics pixels, no ad trackers.
What we store
- Generated analyses live in a short-lived in-memory cache (currently 24 hours) so a repeated visit to
/u/<username>doesn’t re-burn API calls. The cache resets on every deploy. - We do not have a persistent database yet. If you reload the app, your cached analysis may still appear; if the server restarts, it won’t.
- Nothing is published to a public leaderboard without an explicit opt-in flow — which is not live in this version.
What we don’t do
- We don’t sell data. There is no data to sell.
- We don’t email you. We have no email address for you.
- We don’t train models on your analyses.
Removal requests
If you are the owner of a GitHub handle and you want any derived cache entry purged, email hello@githusb.com with the handle. A deploy also wipes the cache.
Third-party links
We link out to github.com for profile lookups and to external share targets (X, Telegram). Those services have their own privacy policies.
Last updated: 2026-04-19. If this page changes materially, the change will be committed to the public repo with a diff.