Privacy and local-only architecture

Invoice files are sensitive. The default scan keeps them in the browser.

The privacy page exists to answer the first enterprise objection: what happens to invoice rows, master data, and exported reports?

Local-only scan path

Simple enough to explain before a client uploads anything.

CSV files Browser worker Local rule engine Report/export Manual wipe / auto-wipe

Privacy proof

The default scanner is intentionally quiet.

No invoice data uploaded by default

CSV parsing and analysis run in the browser. The zero-cost MVP does not require a server-side scan.

No invoice rows in LocalStorage

Preferences may be stored, but raw invoice rows are not persisted in LocalStorage.

Manual wipe and 15-minute auto-wipe

The scan page includes a wipe control and is designed around short-lived local scan sessions.

App shell cached, invoice files not cached

The service worker can cache static assets for reliability. Client invoice files are not cached as app assets.

Enterprise option

Private deployment remains available for firms that prohibit browser tools.

Private hosting

Run in a controlled environment.

For large firms, BPOs, or regulated groups with procurement constraints.

Custom rules

Keep rule packs tenant-specific.

Add ERP-specific mappings, branches, entities, and private exception rules.

Audit posture

Report what was checked.

Timestamp, file names, row counts, deterministic row hashes, and rule-pack versions.

Run the sample scan without uploading client data.