About this project

Archive Suite is a monorepo of three native macOS applications aimed at historians and archivists working through tens of thousands of scanned pages. Each app can run on its own, but they share a documented tag/PDF contract (SPEC/tag-format.md) and are designed to be used together. Archive Processor handles capture, OCR and tagging. It photographs or imports document images, runs OCR, and writes a tagged two-page PDF containing the original image plus selectable OCR text, with Finder tags for subject, date, priority, color and read-state. It includes iPhone and Android capture companions (Live Capture) and guided bring-your-own Gemini or Mistral OCR keys. Archive Reader covers finding, reading and triage. It discovers files by walking the archive folder directly rather than relying on Spotlight, so a stale index cannot hide files; the README cites roughly 100k PDFs in about 10 seconds with an instant warm start from a local cache kept live by FSEvents. It offers chronological navigation, filtering by subject, priority and read-state, full-text search, a two-up image plus OCR viewer, safe tag editing and one-keystroke Read/Unread triage. Archive Notes provides provenance-first note-taking: notes and extracts keep a durable link back to the exact source page that survives moving the computer, plus Zotero references, folders, smart folders, replication and full-text search, designed for 100k notes and 2M words. The README states an irreplaceable-files guarantee: Archive Reader never deletes, moves, renames or alters a file's bytes or location; the only change it makes is to macOS Finder tags, routed through a single audited writer as a deliberate, verified, reversible action. Installation is currently build-from-source only; earlier DMGs were private and withdrawn, and no public release is posted yet. Builds use XcodeGen (project.yml is authoritative, the .xcodeproj is generated and gitignored), with a launch.sh dispatcher for reader, processor or notes, requiring macOS 14+, Xcode 16 and xcodegen via Homebrew. A test-smoke.sh dispatcher runs ArchiveCore unit tests headlessly and free, with optional app lanes; the Processor lane makes two small OCR requests. Live Capture's LAN and USB transports need no setup, while the optional Google Drive relay for restricted Wi-Fi requires a user-created OAuth client bound to the app's package name and signing key. The project is described as a working tool built for one historian's own archive, published for possible reuse rather than as a supported product, maintained by an AI agent with no CI. Known limitations are recorded in each app's KNOWN_ISSUES.md, including an accepted, unfixed weakness in the Live Capture LAN transport. Licensed under MIT.