Khalid Abdi
e8f3ed9ffe
`npm run lint` reported 99 errors and 26 warnings across 65 files and had presumably been failing for a while — next.config.ts sets eslint.ignoreDuringBuilds, so the build never surfaced it. 68 were in vendored code: components/charts and components/ai-elements, pulled from upstream registries. Re-linting those reports upstream's style back at us, and "fixing" them means diverging and eating conflicts on every update — ai-elements already has exactly this carve-out on the TypeScript side via ignoreBuildErrors. Ignore both, plus the two registry files in components/ui (carousel publishes its api from an effect; the sidebar skeleton picks a random width). The other 31 were ours, nearly all react-hooks/set-state-in-effect on the same shape: an effect that re-seeds form state when a dialog opens or a selection changes. Moved to render-phase adjustment, which is both what React recommends and a real fix — the effect version paints one frame of the *previous* record's values before correcting itself. Two carried sharper bugs: the employee dialog could keep a typed password across a switch to another member, and use-wallet-sync could carry `linked` over to a newly-selected patient, briefly offering to push a record to someone else's wallet. The rest: useIsMobile and speech-support detection become useSyncExternalStore (correct on first paint, no mount flash); refs mirroring state are written in effects rather than during render; the care-team fetch moves into its effect behind a reload key, dropping an exhaustive-deps suppression. Two effects in chat-panel keep the rule disabled with a reason. Both are what effects are for: draining the queued-message buffer when the transport goes idle, and resolving ?thread from the URL — the latter mints an id with nanoid(), so moving it into render would just trade this error for a purity one. Also removes a dead /explore-era import and unused directives found on the way. lint now exits 0, with the ai-elements tsc carve-out unchanged. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
temetro
An open-source AI middleman between clinicians and patient data.
Clinicians ask in plain language; temetro retrieves and organizes patient information as rich record cards — backed by a patient-owned data model.
What is temetro?
temetro is a clinical tool that puts a natural-language AI chat between clinicians and patient records. Instead of clicking through tabs, a clinician asks for what they need and temetro returns organized record cards — vitals, labs, medications, problems, encounters — with trend sparklines and detail views.
Its distinguishing idea is a patient-owned data model. Instead of (or alongside) living only in a doctor's database, a patient's record can live on the patient's own device. When a clinician adds or changes data they sign it (blockchain-style); the change is written to the patient's record and cannot be modified until the patient approves it through a companion wallet app. temetro can also read existing patient databases and present them in the same card UI.
"Decentralization" here means keys and data live on the patient's device and the relay only ever forwards ciphertext — it is not a literal blockchain. Records are off-chain, which is what lets a temporary share be deleted.
Features
- 🗂️ AI chat over patient records —
/patient <file#>(or natural language) renders the record as cards with sparklines and detail dialogs. - 🔐 Auth & multi-tenant clinics — email/password + staff usernames, organizations, and role-based access (owner / admin / doctor / reception / pharmacy / lab) via Better Auth.
- 🩺 Full clinical surface — patients, appointments, prescriptions, tasks, doctor's notes, pharmacy inventory, analytics, an activity audit log, real-time staff messaging, and notifications.
- 📲 Patient wallet & encrypted share — a companion app holds the record encrypted on-device; clinics import records over an end-to-end encrypted, patient-approved relay (with optional auto-deleting temporary shares).
- 🌐 Works across the clinic LAN — open it on the server or from any
department's computer at
http://<server-IP>:3000; no per-machine config. - 🔄 Self-update awareness — the app tells admins when a newer release is out and how to update.
Quick start (Docker)
The fastest path uses the prebuilt images published to Docker Hub. From the
backend/ directory (it holds the Compose file):
cd backend
docker compose pull # fetch the latest published images
docker compose up -d # start Postgres + backend + frontend
No .env or secret setup is required — the backend generates and persists any
missing secrets on first start. Then open:
- Frontend → http://localhost:3000
- Backend → http://localhost:4000 (health:
GET /health)
Prefer to build from source (for development)? Use docker compose up --build instead. Migrations apply automatically on backend start.
Port conflict? If host ports
5432,4000or3000are already in use, setPOSTGRES_PORT,BACKEND_PORTand/orFRONTEND_PORT(e.g.5433/4001/3001) inbackend/.env. The services still talk to each other on their internal ports (db:5432,backend:4000); only the published host ports change.
Access from other computers (hospital LAN)
temetro figures out the backend address from the host you open it on, so other
departments can simply visit http://<server-LAN-IP>:3000 — no rebuild
needed. Settings → About & updates shows the exact shareable address. The
server's firewall must allow ports 3000/4000.
Updating
docker compose pull && docker compose up -d
The app surfaces a notification when a newer release exists. See
CHANGELOG.md for what changed and RELEASING.md
for how releases are built and published.
Run without Docker
Each app runs independently (you'll need a local Postgres in DATABASE_URL):
# backend
cd backend && npm install && cp .env.example .env
npm run db:migrate && npm run dev # http://localhost:4000
# frontend (second terminal)
cd frontend && npm install && npm run dev # http://localhost:3000
With
SMTP_HOSTunset, verification / reset / invitation emails are printed to the backend console — no email setup needed for local development.
Monorepo layout
Two independent apps live side by side, each with its own package.json and
CLAUDE.md:
frontend/— Next.js 16 product app (the AI chat UI).backend/— Express 5 + Postgres API (Drizzle ORM + Better Auth). Seebackend/README.mdfor the API reference.
The patient wallet app and the marketing landing page live in their own separate repositories.
Tech stack
- Frontend: Next.js 16 (App Router) · React 19 · TypeScript · Tailwind CSS v4 · i18next · Socket.io client.
- Backend: Node ≥ 20 · TypeScript (ESM) · Express 5 · Postgres · Drizzle ORM · Better Auth (email/password + organizations/RBAC) · Socket.io · AI SDK.
Contributing
Issues and pull requests are welcome. Please keep each PR focused on one logical
change and run the type checks (npm run typecheck in backend/, npx tsc --noEmit in frontend/) before opening it. Bug reports and feature requests
have issue templates.
License
MIT.