🩺 MediMind — AI‑Powered Prescription + Patient Management (Full‑Stack Monorepo)

MediMind is a full‑stack clinic workflow app for doctors to:

• register/login (JWT)
• manage patients
• generate AI‑assisted prescriptions (Gemini via a FastAPI service)
• save prescriptions (with nested prescription items) in the backend

Medical disclaimer: This project can generate medical‑sounding content. It is a demo/prototype and must not be used for real clinical decisions without qualified review and safety controls.

---

Monorepo structure

MediMind/
├── MediMind-Frontend/   # Next.js (App Router) UI
├── MediMind-Backend/    # Django + DRF API (JWT auth)
└── MediMind-AI/         # FastAPI service that calls Gemini and returns JSON

---

Tech stack (as implemented)

Frontend

• Next.js 15.x (React 19)
• TypeScript
• Tailwind CSS
• react-hot-toast

Backend

• Django 5.2
• Django REST Framework
• JWT via djangorestframework-simplejwt
• DB config via dj-database-url
• CORS via django-cors-headers
• Static files via WhiteNoise

AI service

• FastAPI + Uvicorn
• LangChain + langchain-google-genai
• Pydantic v2 response schema parsing

---

How the system fits together

1. Frontend authenticates with Backend (/users/login/) and stores the JWT.
2. Frontend fetches doctor profile (/users/me/) and doctor’s patients (/patients/doc<id>/).
3. For AI generation, Frontend calls AI service (POST /generate_prescription).
4. After review/edits, Frontend saves the prescription to Backend (POST /prescriptions/).

---

Quickstart (local dev)

You’ll run 3 servers in 3 terminals:

• Backend: Django API (recommended: http://127.0.0.1:8000)
• AI service: FastAPI (recommended: http://127.0.0.1:8001 to avoid a port clash)
• Frontend: Next.js (default: http://localhost:3000)

Prerequisites

• Node.js 18+ (recommended: 20+)
• Python 3.10+ (recommended: 3.11/3.12)

---

1) Backend setup (Django + DRF)

Install

cd MediMind-Backend
python3 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
pip install -r requirements.txt

Environment

Create MediMind-Backend/.env:

cat > .env << 'EOF'
DJANGO_SECRET_KEY=dev-secret
DJANGO_DEBUG=True

# IMPORTANT:
# The backend currently uses dj-database-url and expects DATABASE_URL.
# If you want SQLite locally, see the note below.
DATABASE_URL=
EOF

Database note (important)

In MediMind-Backend/MediMind/settings.py, the default database is loaded from DATABASE_URL and sets ssl_require=True.

• For production Postgres this is great.
• For local development, this may break typical local Postgres URLs (no SSL).
• A SQLite configuration block exists in the file but is commented out.

If you want an out‑of‑the‑box local run with SQLite, uncomment the SQLite DATABASES = {...} block and comment out the Postgres dj_database_url.config(...) block.

Migrate + run

python manage.py migrate
python manage.py runserver

Backend will be at:

• http://127.0.0.1:8000/

Admin (if you create a superuser):

• http://127.0.0.1:8000/nitish/

---

2) AI service setup (FastAPI + Gemini)

Install

cd MediMind-AI
python3 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
pip install -r requirements.txt

Environment

Create MediMind-AI/.env:

cat > .env << 'EOF'
GEMINI_API_KEY=your-gemini-key
GEMINI_MODEL=gemini-2.5-flash
EOF

Run

The service defaults to port 8000 in many examples, so run it on 8001 locally to avoid colliding with Django:

uvicorn app.main:app --reload --port 8001

AI service will be at:

• http://127.0.0.1:8001/
• Docs: http://127.0.0.1:8001/docs

---

3) Frontend setup (Next.js)

Install

cd MediMind-Frontend
npm install

Environment

Create MediMind-Frontend/.env.local:

cat > .env.local << 'EOF'
NEXT_PUBLIC_BASE_URL=http://127.0.0.1:8000
NEXT_PUBLIC_AI_URL=http://127.0.0.1:8001
EOF

Run

npm run dev

Frontend will be at:

• http://localhost:3000/

---

API reference (actual implemented routes)

Backend (Django)

Base URL: NEXT_PUBLIC_BASE_URL (example: http://127.0.0.1:8000)

Auth / Users

• POST /users/register/ — register doctor (public)
• POST /users/login/ — JWT login (public, SimpleJWT)
• POST /users/refresh/ — refresh JWT (public, SimpleJWT)
• GET /users/me/ — current user profile (auth)

Auth header for protected routes:

• Authorization: Bearer <access_token>

Patients

• GET /patients/?search=<text> — list patients (auth)
• POST /patients/ — create patient (auth)
• GET /patients/doc<doctor_id>/ — list patients by doctor (auth)
• GET /patients/<id>/ — patient detail (auth)
• PUT/PATCH /patients/<id>/ — update (auth)
• DELETE /patients/<id>/ — delete (auth)

Prescriptions

• GET /prescriptions/ — list prescriptions (auth)
• POST /prescriptions/ — create prescription (auth)

Prescription create expects nested prescription_items.

Example request body:

{
  "patient": 1,
  "symptoms": "fever, sore throat",
  "diagnosis": "viral pharyngitis",
  "notes": "hydration, rest",
  "prescription_items": [
    {"medicine": "Paracetamol", "dosage": "500mg", "instructions": "Twice daily after meals"}
  ]
}

---

AI service (FastAPI)

Base URL: NEXT_PUBLIC_AI_URL (example: http://127.0.0.1:8001)

• POST /generate_prescription

Request body:

{
  "name": "Alex Doe",
  "age": 34,
  "gender": "male",
  "allergies": "penicillin",
  "medical_history": "asthma",
  "symptoms": "fever, sore throat, cough"
}

Response:

{
  "diagnosis": "...",
  "notes": "...",
  "prescription_items": [
    {"medicine": "...", "dosage": "...", "instructions": "..."}
  ]
}

---

Frontend scripts

From MediMind-Frontend/package.json:

• npm run dev
• npm run build
• npm run start
• npm run lint

---

Common issues / troubleshooting

Backend fails to start: database misconfiguration

If DATABASE_URL is empty/invalid, Django may fail to connect.

Fix options:

• Provide a valid Postgres DATABASE_URL (recommended for production), or
• Switch to SQLite for local dev by uncommenting the SQLite DATABASES block in MediMind-Backend/MediMind/settings.py.

Port conflict (8000 already in use)

Both Django and FastAPI commonly run on 8000.

• Keep Django on 8000
• Run AI on 8001 (uvicorn ... --port 8001)
• Set NEXT_PUBLIC_AI_URL=http://127.0.0.1:8001

CORS

• Backend currently allows all origins via CORS_ALLOW_ALL_ORIGINS = True.
• AI service also allows all origins (allow_origins=['*']).

This is convenient for development but not recommended for production.

---

Security + privacy notes

• JWT access tokens are required for most backend routes (IsAuthenticated is global by default).
• Don’t store real patient data in dev/demo environments.
• Lock down CORS, secrets, and database credentials before deploying.

---

License

MIT — see LICENSE.

---

Medical disclaimer

MediMind is a software prototype designed to assist clinicians. AI output may be incomplete, incorrect, or unsafe.

• Always validate outputs clinically.
• Do not rely on this system as a substitute for medical judgment.
• Ensure compliance with privacy regulations and internal policies.