Privacy policy
PAM is built on a metadata-only principle. Here's exactly what we see, what we don't, and how your data is handled.
PAM is a document index, not a document reader. When you connect a folder, PAM reads filenames, folder paths, and file metadata (size, creation date, modification date). That's it. It never opens files, never reads their contents, and never stores anything from inside a document.
84% of documents are classified from metadata alone — no content access required.
Some documents can't be identified from a filename alone. For these, PAM offers optional deep analysis — reading the first 500 characters of a document to improve classification confidence. This is always opt-in and requires your explicit permission.
Any content read during deep analysis is ephemeral. It is used only in the moment to classify the document and is never stored, indexed, or committed to the database. Once classification is complete, the content is discarded.
PAM uses Claude (Anthropic's AI) to classify documents. When classification runs, it sends the entity name (e.g., "2018 Toyota Camry") and the document filename to the Claude API. It never sends document contents.
Anthropic's API terms apply to data sent for classification. We don't use classification data to train models, and we don't store API responses beyond the classification result.
PAM lets you generate secure share links for entities — useful for sharing relevant documents with an accountant, attorney, or property manager. Share links use 16-byte cryptographically random tokens and can be set to expire.
By default, share links hide your personal identifiers (account numbers, ID numbers) and private notes. These fields are only included if you explicitly grant visibility when creating the link. The principle: safe by default, permissive only when you decide.
If you join the beta waitlist, we collect your name, email address, and the persona you identify with (e.g., Owner, Steward, Executor). This information is used only to send product updates and coordinate beta access. We do not sell it, share it with third parties, or use it for advertising.
You can ask to be removed from the waitlist at any time by emailing us.
The PAM marketing site uses no analytics cookies, no tracking pixels, and no third-party data sharing. We don't fingerprint visitors or run retargeting campaigns. This page loads Google Fonts — that's the only third-party request the site makes.
The PAM product app (when in beta) may use lightweight, privacy-respecting analytics to understand feature usage. Any such data is aggregate, not linked to individuals, and is never sold.
PAM runs on your local machine and stores data in a SQLite database you own. There is no cloud sync, no remote data store, and no third-party data pipeline. Your document index lives where you run PAM — and only there.
If you want to delete your data, delete the database file. No account to close, no deletion request required.
Questions about this policy or how PAM handles your data? Reach us at hello@pam.wherewithal.studio.