🚧 Work in progress. Actively evolving, expect frequent changes to scope, UI and data model. Not production-ready (see Security notes below).
I built this as a functional prototype to validate the business logic before putting any real money into a full product. It's a fictional hair/beauty diagnostic and booking app, and it covers: client accounts (login/password), a post-login category picker (Hair for Women, Hair for Men, Makeup, Skincare, all sharing one account and one cart), hair anamnesis, harmony assessment (visagism + personal color analysis), a full protocol menu with no budget gating, curated Makeup and Skincare catalogs grounded in researched 2026 K-beauty trends, an optional budget-capped quote, a technical data sheet per product (usage, ingredients, consumption plan, embedded video tutorial), a lightweight gamification layer (points, progress bar, achievement badges), appointment booking with a partner professional, and product orders (no real payment processing).
Built with zero external dependencies, a single Node.js HTTP server, in-memory sessions, and a JSON file as the data store, so it runs anywhere Node is installed, with no install step and no network dependency.
No npm install needed, it only uses Node.js built-in libraries (no external dependencies, by design, to avoid any network requirement while testing).
node server.jsThen open http://localhost:3000 in your browser. Create an account (email + password, 6 characters minimum) and try the flow.
Requirement: Node.js 18 or later (node -v to check).
| Hair anamnesis & gamification bar | Makeup catalog (real, researched 2026 trends) |
|---|---|
 |
 |
- Create a new GitHub repository (e.g.
j8-app) and push this content to it (dragging files into GitHub's "Add file → Upload files" page works too, no command line required). - Go to
https://render.com/deploy?repo=<your GitHub repo URL>, Render reads therender.yamlalready included here and proposes the configuration automatically (free Node web service, start command pre-filled). - Confirm and wait ~2-3 minutes. You'll get a link like
https://j8-app-xxxx.onrender.com.
Important about Render's free tier: storage isn't persistent, data/db.json (accounts and saved records) can be wiped whenever the service restarts or redeploys. Fine for letting friends test the flow for a few days; not yet a foundation for storing real client data long-term (a persistent disk or a proper database would be the next step for that).
- Sign up / log in / log out (password hashed with
scrypt, never stored in plain text) - Post-login category picker (Hair for Women, Hair for Men, Makeup, Skincare), all four route to the same account and the same shopping cart; switch between them anytime from the pill bar without losing anything
- Hair anamnesis using clinical language (main concern, gender, facial geometry, skin undertone, color direction), gender is pre-filled from the chosen category but stays editable
- Harmony assessment (visagism + personal color analysis), always presented with an alternative, never a single verdict
- Full protocol menu, every phase and tier (Essential/Clinical) freely selectable, with no budget cap filtering the options
- Single placeholder brand ("HANA LAB", fictional, Korean-beauty-inspired) applied consistently across the product catalog, to keep the demo coherent until a real supplier is chosen
- Makeup catalog (5 curated items) and Skincare catalog (Top 5 Korean skincare essentials), each grounded in real, researched 2026 K-beauty trend reporting (cited in commit messages), not invented; both are a curated starter kit rather than a full diagnostic quiz like Hair has
- Custom hand-drawn SVG product illustrations (dropper bottle, jar, cushion compact, tube, ampoule, sheet-mask sachet, cream pot) in the HANA LAB gradient style, replacing the earlier camera-emoji placeholder, this environment couldn't reach stock-photo CDNs to fetch real photographs, so illustrations were the practical path to "more visual, not a placeholder icon"
- Embedded video tutorials (real public YouTube videos, found via research, double cleansing, sheet masks, cushion foundation application, hair oiling/scalp massage, a 10-step glass-skin routine overview) on the relevant product cards
- Lightweight gamification layer, points counter, progress bar toward a complete protocol, and unlockable achievement badges (cosmetic only, doesn't change pricing)
- Per-product technical sheet: description, step-by-step usage, key ingredients, benefits, contraindications ("do not use if"), a consumption plan (estimated yield until the next purchase), a "2026 trend" callout, and (where available) a video tutorial
- Budget-capped quote, optional, only activates if the client asks for it, at the end of the flow
- Client record, assessment history saved per account
- Shopping cart persists across page reloads, tab navigation and re-login (
localStorage), cleared only on logout - Booking calendar with 3 sample professionals, available slots, booking and cancellation
- Product orders (resale), logged as "pending", no real payment gateway
- Automated Python/Selenium/Pytest test suite covering the flows above (see
tests/, test case matrix intests/TEST_PLAN.md), run automatically on every push via GitHub Actions (.github/workflows/tests.yml), see the "Tests" badge at the top of this file for the current status
A long list of additional features accumulated over the course of requests. To keep this prototype testable (and avoid a project that never ships), they're recorded here as next steps rather than built blindly:
- A real diagnostic quiz for Makeup and Skincare (today they're a curated product grid, not an anamnesis flow like Hair's)
- Admin dashboard (sales KPIs, profit/loss, inventory, stock shortages/surplus, alerts for difficult clients/complaints)
- Stock management with pricing, shortages and surplus
- Pickup points
- Installment payment plans
- Apple Pay / Google Pay / real payment checkout
- Insurance (unclear which type, product, service, liability, needs clarification before designing)
- Real AI engine (today the harmony/protocol responses are deterministic rules, not calls to a language model; a future version could plug in an LLM for that)
- Real product photography and a real supplier relationship (currently illustrated, under one fictional placeholder brand, "HANA LAB", see
tests/TEST_PLAN.mdand commit history for the trend research this content is grounded in)
Each of these is a real chunk of work (most involve integrating an external service, or real money moving through the system), worth treating as a separate request rather than another line of code in the same file.
- Sessions live in server memory, restarting the process ends all active sessions (fine for testing, not for production)
- The database is a local JSON file (
data/db.json), no encryption at rest, no automatic backup - No HTTPS (runs on
http://localhost), never expose this server directly to the internet without TLS - No login rate limiting, add this before any use beyond localhost
- Handling real client data in production would also require a full LGPD/GDPR and Anvisa/INFARMED compliance review (data retention, consent, right to erasure), not yet addressed in this prototype
Every push to main (and every pull request) runs the full Selenium/Pytest suite against a freshly started instance of the server on GitHub-hosted runners, see .github/workflows/tests.yml. This runs the real suite with real internet access (installing Selenium, Pytest and Chrome from scratch each run), independent of any local environment quirks.
MIT, see the LICENSE file for the full text.