MaatriSakhi

A conversational pre-consultation assistant designed to help patients organize their information before meeting a healthcare professional.

Comment

🌸 MaatriSakhi — Preconception Care Assistant

MaatriSakhi Logo

AI-Assisted Preconception Care & Clinical Screening

A conversational pre-consultation assistant designed to help patients organize their information before meeting a healthcare professional.

Live URL: https://maatrisakhi.onrender.com/

Demo: https://youtu.be/oe1OXRG6B8c?si=8Ibr8IpfuBgEiKTu

📖 Primary Clinical References

This application is strictly grounded in the official medical guidelines of the Federation of Obstetric and Gynaecological Societies of India (FOGSI):

  1. Preconception Care E-Booklet, Book 1: "Preconception: Building the Foundation for a Healthy Mother & a Healthy Future"
    • Editor-in-Chief: Dr. Bhaskar Pal (President, FOGSI)
    • Editors: Dr. Suvarna Khadilkar, Dr. Priti Kumar, Dr. Poonam Goyal (Chairperson, Safe Motherhood Committee)
  2. FOGSI Preconception Care — Complete Clinician Checklist
    • 12 structured clinical protocols covering pregnancy intention, medical disorders, previous obstetric complications, medications, genetic carrier risks, immunization, environmental toxins, lifestyle, and mental health.

🛡️ Strict Medical Safety Guardrails

  • Pre-Consultation Assistant Only: The application does NOT diagnose medical conditions, prescribe drugs, or substitute for a qualified doctor.
  • Topics Flagged for Clinician Review: Uses neutral reminders with status indicators (🟢 No issue reported, 🟡 Review with clinician, 🔴 Important clinician attention) rather than synthetic, alarmist "risk scores".
  • Medication Review Without Hazardous Discontinuation: Never independently tells a patient to stop medicines (particularly psychotropic, antihypertensive, or anti-epileptic drugs), avoiding dangerous withdrawal relapses.
  • Evidence-Based Folic Acid Recommendations: Explains FOGSI preconception folic acid guidelines:
    • Low Risk: 400–800 μg/day starting at least 1 month before conception.
    • Higher Risk (Diabetes, Epilepsy, prior Neural Tube Defect): 4–5 mg/day starting 1–3 months before conception.

✨ Key Features

  1. One Question at a Time Conversational Experience:
    • Clean, friendly digital health assistant avoiding overwhelming hospital forms.
    • Large, accessible Green YES (#16a34a), Red NO (#dc2626), and Slate NOT SURE (#475569) buttons.
  2. Voice-First & Text-To-Speech:
    • Web Speech API integration (SpeechRecognition) for hands-free answering in native languages.
    • Built-in Text-To-Speech (SpeechSynthesis) with play/stop controls and an auto-read toggle.
  3. 8 Indian Languages Supported:
    • English, हिंदी (Hindi), తెలుగు (Telugu), தமிழ் (Tamil), ಕನ್ನಡ (Kannada), മലയാളം (Malayalam), বাংলা (Bengali), मराठी (Marathi).
    • Medical terms annotated with parenthetical English names for clinical clarity.
  4. Smart Branching Logic:
    • Adaptive follow-up questions triggered only when relevant (e.g. asking for HbA1c and duration only if diabetes is reported; skipping obstetric complications if nulliparous).
  5. Preconception Consultation Summary & Clinician Dashboard:
    • Categorized pre-visit summary ready before the patient enters the consultation chamber.
    • Rule-based flags highlighting critical clinical review items.
    • Direct FOGSI booklet page and checklist citations.
    • Doctor's consultation notes and sign-off checkbox ("Checklist completed and discussed with the couple").
    • Print & PDF export layout optimized for hospital records.
  6. One-Click Realistic Demo Patient:
    • Instant simulation of a patient case (Ananya Sharma, 28, planning conception, mild hypothyroidism on levothyroxine, not yet taking folic acid, partner smoking) to test the entire end-to-end workflow in seconds.
  7. One-Click Mother Demo Account (Pregnancy + Child flows):
    • Email: [email protected] · Password: mother123 (same convention as the doctor demo: [email protected] / doctor123).
    • Pre-seeded with consent given + a week-38 pregnancy (high BP + gestational diabetes + doctor limits, visit in 7 days) so both trackers and the Child Health Card creation flow are immediately explorable. Works online (Postgres) and offline (local demo store).
    • The doctor login screen has a matching one-click demo ([email protected] / doctor123).

🤱 Pregnancy Companion Mode (In Development)

Phase 1: Storage, Login, Consent - Database infrastructure implemented:

  • PostgreSQL database with 4 new tables (mothers, pregnancies, entries, visits)
  • JSON schema validation for all data
  • DPDP Act 2023 consent compliance
  • Medical thresholds marked as "placeholders, to be confirmed by clinicians"
  • Render.com PostgreSQL deployment ready

Phase 2-11: Mode Switch & Pregnancy Features - Not yet implemented.

  • New "I'm pregnant" mode will run alongside existing "Planning a pregnancy" flow
  • Uses separate database tables to maintain zero impact on existing preconception data
  • All 8 languages and voice support will carry over
  • Existing preconception app flow completely unchanged

Safety: FOGSI guidelines only, pre-consultation only, no diagnoses, no prescriptions, never advise stopping medication. Out-of-range readings show "Contact your doctor" only.

🚀 Quick Start

Prerequisites

  • Node.js 18+
  • npm or yarn

Installation

# Clone the repository
git clone https://github.com/alwaysalearner1234/MaatriSakhi.git
cd MaatriSakhi

# Install dependencies
npm install

# Start development server
npm run dev

Open http://localhost:5173/ in your browser.

Production Build

npm run build
npm run preview

📂 Project Architecture

MaatriSakhi/
├── public/
├── src/
│   ├── components/
│   │   ├── AboutModal.jsx          # FOGSI guidelines & safety modal
│   │   ├── ChatInterface.jsx       # Conversational chatbot with voice/TTS
│   │   ├── DoctorDashboard.jsx     # Clinician pre-visit summary & review flags
│   │   ├── LanguageSelector.jsx    # 8-language selection cards
│   │   ├── PatientReview.jsx       # Pre-submission review & answer editing
│   │   └── WelcomeScreen.jsx       # Reassuring welcome & safety disclaimer
│   ├── data/
│   │   ├── demoData.js             # Realistic simulated patient profiles
│   │   ├── fogsiQuestions.js       # Structured FOGSI questions & flag logic
│   │   └── translations.js         # Multi-language dictionaries
│   ├── utils/
│   │   └── speechUtils.js          # Web Speech API recognition & TTS synthesis
│   ├── App.jsx                     # Core application coordinator
│   ├── index.css                   # Responsive FOGSI clinical design system
│   └── main.jsx
├── index.html
├── package.json
└── vite.config.js

🚀 Render Deployment Instructions

Prerequisites

  • Render.com account
  • Supabase project (free tier OK)

1. Get Supabase Connection String

  1. Go to Supabase dashboard → Settings → Database
  2. Note the "Connection string" for the "Postgres pooler (session mode)"
  3. It will look like: postgresql://postgres:[email protected]:6543/postgres

2. Add Python Web Service

  1. Go to Render dashboard → New Web Service
  2. Select "Deploy from a Git Repository"
  3. Repository: your MaatriSakhi repo
  4. Build Command: pip install -r requirements.txt
  5. Start Command: uvicorn api:app --host 0.0.0.0 --port $PORT
  6. Service Type: Web Service

3. Environment Variables

Add these to the Web Service:

DATABASE_URL=postgresql://postgres:[email protected]:6543/postgres?sslmode=require
SECRET_KEY=your-super-secret-key-for-jwt-tokens
NODE_ENV=production
ALLOWED_ORIGINS=http://localhost:5173

4. requirements.txt

Create requirements.txt at the project root:

fastapi
uvicorn[standard]
asyncpg
bcrypt
PyJWT
pydantic

5. Deploy

  • Push code to git
  • Render will auto-detect and build
  • The Static Site serves the React frontend on port 80
  • The Python Web Service runs the API on port $PORT

6. Verify

  • Frontend: https://maatrisakhi.onrender.com/ (React SPA)
  • API: https://your-service-name.onrender.com/docs (FastAPI auto-docs)
  • Health: https://your-service-name.onrender.com/health

📜 License & Medical Disclaimer

Designed for clinical workflow assistance. The tool collects information for healthcare consultations and does not replace medical advice, clinical diagnosis, or patient-physician evaluation.

Built With

Share this project:

Updates

Submission history