MedTalk is an AI-powered healthcare assistant that helps users understand their prescriptions. Upload a prescription image, and MedTalk extracts medicine details, provides a summary, and enables a conversational chat to answer follow-up questions in simple language.
- Upload Prescription Images: Snap or upload a photo of your prescription.
- OCR Extraction: Automatically reads and extracts medicine names and instructions.
- AI Chat: Ask questions about your medicines, usage, side effects, and more.
- Prescription Summary: Get a concise, patient-friendly summary of your prescription.
- Verify Medicines: Quickly check if your purchased medicines match the prescription.
- Responsive UI: Built with Tailwind CSS for a seamless experience on any device.
- Frontend: React, Tailwind CSS
- Backend: Node.js, Express, Redis (for caching)
- OCR: Puter.ai
- AI: Google Gemma (for chat and summaries)
git clone https://github.com/tilakjain619/MedTalk.git
cd MedTalk# Install frontend dependencies
cd client
npm install
# Install backend dependencies
cd ../server
npm install- Create a
.envfile in bothclientandserverfolders as needed. - For example, in
client/.env:VITE_BACKEND_URL=http://localhost:5000 - In
server/.env, add your Openrouter API key and redis credentials:OPENROUTER_API_KEY=your_api_key REDIS_HOST=your_redis_host REDIS_PORT=redis_port REDIS_USERNAME=default REDIS_PASSWORD=your_redis_password
# Start backend
cd server
node app.js
# In a new terminal, start frontend
cd ../client
npm run dev- Upload a prescription image on the main page.
- Wait for the app to extract and summarize the prescription.
- Use the chat to ask follow-up questions like:
- "What is this medicine for?"
- "Are there any side effects?"
- "How should I take these medicines?"
- Use the Verify medicines feature to quickly verify if you have bought the right medicines.
For questions or support, contact @tilakjain619.