Privacy
Last updated 2026-07-20
The free tool runs entirely in your browser
When you paste or upload a file list, it is parsed and analyzed with JavaScript on your own machine. It is never uploaded, and we never see it.
The paid API sees file paths and sizes — never file contents
When you connect the GitHub Action or pre-commit hook, it sends us a list of your file paths and their byte sizes — plus the keep: lines from a committed .contextpruner file, if you use one — so we can generate all nine config files, the advisory instruction files and the enforced ignore files alike, and send them back. We never receive the contents of your files. We store only aggregate counts for each run — how many files, total bytes, how much was pruned, and which tool called — tied to your account so your usage dashboard works. We do not store your file paths, your declarations, or your file contents.
The runtime filter shares only opt-in aggregate counts
The paid contextpruner serve command — the runtime search filter — runs entirely on your machine. The first time you run it, and only in an interactive terminal, it asks once whether to share two running totals: how many files its filter has assessed and how many bytes of junk it has kept out of your agent's context. If you say yes, those two numbers are all that is ever sent — no paths, no file contents, no per-file data — and we store them as anonymous aggregate totals, not tied to your account. If you say no, which is the default, or run it in a non-interactive shell, nothing is ever sent. Everything else the filter does stays on your machine.
Signing in
You sign in with GitHub. We ask GitHub only to identify you — we never request access to your repositories. We store your account identifier and, if GitHub shares it, your email address.
Payments
Billing is handled by Stripe. Your card details go straight to Stripe and never touch our servers. We store your Stripe customer and subscription identifiers and your subscription status.
API keys
Your key is shown once, at creation. We store only a SHA-256 hash of it — we can't recover the key, and neither can anyone who reads our database.
Analytics
We use Vercel Web Analytics, which is cookieless. We record anonymous, aggregate events — that a manifest was analyzed, the sample was loaded, the savings counter was viewed, an upgrade was clicked, checkout was started — and page views. These never include your file paths, file contents, or any personal information.
Staying signed in
Your session is kept in your browser's local storage so you don't have to log in every visit. We don't use tracking cookies.
Your data
Deleting your account removes your account, keys, and usage records. To request deletion or ask a question, email support@contextpruner.app.