bilalkaradeniz
← Back to all work

2026 — now · Creator, full-stack

Foodlex

Client tracking and AI-assisted diet planning for dietitians.

Seven-day plan drafts, approved by the dietitian
Branded PDF plans, sent by email
Health data encrypted, servers in the EU
Stack
Next.js · React · TypeScript · NextAuth · FastAPI · Python · SQLAlchemy · PostgreSQL · LLM APIs
foodlex.co ↗
sessionsasyncplan jobsDietitianNext.jsFastAPISQLAlchemyPostgreSQLNextAuthLLM APIssessionsasyncplan jobsDietitianNext.jsFastAPISQLAlchemyPostgreSQLNextAuthLLM APIs

Problem

Dietitians juggle spreadsheets, messaging apps and paper forms, re-collect the same client details at every visit, and write each diet plan from scratch.

System

A Next.js interface with NextAuth sessions over an async FastAPI and PostgreSQL backend, with Alembic migrations. Diet-plan generation runs as tracked jobs against LLM APIs, so usage is metered and every draft is reviewed by the dietitian.

Foodlex website
foodlex.co

Hard parts

  • Modelling clients, foods, recipes, plans and appointments so the daily workflow stays fast.
  • Privacy-minded handling of health data, including export and full account closure.
  • Tested on both ends with pytest and Vitest.

Result

Live at foodlex.co with plans for 10, 50 and 200 clients and a 14-day free trial. Clients move over from Excel or CSV, health and contact data are encrypted on EU servers, and consent is recorded under KVKK, Türkiye’s data protection law. Built by me.

Next projectVoText