About this project

FED-SPA (Florida Establishment Directory — Spa & Parlor Assurance) is an independent public-records project that publishes a static, once-a-year snapshot of licensed massage establishments in Florida, checked by hand against the Florida Department of Health Medical Quality Assurance (MQA) portal. The README states the current snapshot as of 2026-09-11: 47 licensed records and 44 watchlist entries, release 2026.2. The project is built around three stated constraints: no backend (static files only, no server, API, database or accounts), no third-party code (Python stdlib, Node native crypto, vanilla JS/CSS, androidx core + appcompat, SwiftUI + CryptoKit/CommonCrypto), and an annual refresh cadence rather than a live feed. Data is split into two tiers. The licensed list is public record and ships as plaintext JSON (data/public/licensed.json), described as auditable and forkable, with a CSV mirror generated from the same source. The watchlist covers establishments where a documented MQA search found no license match; it is encrypted at rest using PBKDF2-SHA256 to AES-256-GCM and decrypts in-app after a subscriber enters a shared code. Each watchlist entry records the exact search terms and date in a status_note, and the README stresses that a missing license is a prompt to ask questions, not a verdict. A single dataset is fanned out by admin_scripts/generate_public_files.py to six surfaces: an installable offline-capable PWA, a Chrome Manifest V3 extension with domain badge and popup, an Android app with deep links and a home-screen widget, an Android Auto surface, a watch surface, and a SwiftUI iOS app. Android Auto and watch are deliberately limited to licensed data only. The repository includes admin scripts for the annual workflow (scrape_mqa.py, encrypt_unlicensed.js, validate_schema.py, generate_public_files.py, merge_and_encrypt.sh), documentation for install, build, deployment, legal disclaimers and contribution guidelines, wiki pages covering the platform surface map, crypto envelope spec and refresh log, and ready-to-use AI-assistant prompts. Records carry verified_by, last_checked and data_as_of fields, and corrections are handled through an issue template. Licensing is split: code under MIT, dataset under ODbL 1.0. The README states the project is independent and unaffiliated with the State of Florida or any agency, is not a government service or live feed, and urges re-verification on the MQA portal before any decision.