Poshan is an AI-driven healthcare platform designed to provide accessible and reliable medical assistance, especially for rural communities. The level of personalization being one of the main usp.
Medical Report Analysis: Upload and interpret medical reports for better understanding.
Medication Details: Access comprehensive information about medications, including usage and side effects.
Nutrition Insights: Receive dietary recommendations and insights tailored to individual health needs.
Chatbot Assistance: Engage with an AI-powered chatbot for instant health-related queries.
Multilingual Support: Interact in various regional languages to cater to diverse user bases.
Frontend: Next.js, TypeScript, Tailwind CSS
Backend: Node.js, FastAPI
AI & ML: OpenBioLLM-70B, LangChain, NLTK
Data Storage: Pinecone
APIs & Integrations:REST API, Optical Character Recognition (OCR), Gemini API
Users can interact with the Poshan chatbot via the web interface:
Upload and analyze medical reports.
Inquire about medication details.
Seek nutrition and dietary advice.
Get instant responses to health-related questions.
Progress During Hackathon – Poshan: AI-Powered Healthcare Companion During the course of the hackathon, our team made substantial progress on both the front-end and back-end of Poshan, laying down the foundation for a fully functional AI healthcare assistant tailored to rural needs. ✅ Core Features Implemented: Built a clean, responsive front-end using Next.js, TypeScript, and Tailwind CSS, with a working chatbot UI. Integrated OCR functionality to allow users to upload and analyze medical reports. The system extracts relevant data and passes it to the AI model for interpretation. Connected the chatbot to OpenBioLLM-70B via LangChain, enabling it to respond to medical, medication, and nutrition-related queries. Developed nutrition recommendation logic using custom NLTK pipelines and health parameters. Enabled multilingual support for Hindi and English using prompt-engineered Gemini API for translation and context handling. Set up FastAPI routes and connected them with Node.js APIs for seamless back-and-forth between the front-end and model logic. Initialized Pinecone vector storage for semantic search and context retrieval. 💡 What We’re Excited About: The chatbot not only interprets medical reports but can simplify them in regional languages—a key step toward healthcare inclusivity. Personalized nutrition suggestions based on extracted vitals and conditions, bringing a humanized feel to automated care. 🛠️ What’s Next: Adding report history tracking. Integrating Google maps api to connect user with nearest doctors and further personalization. Polishing the UI/UX consistently for a smoother flow and rural accessibility. Overall, Poshan has progressed from idea to functional prototype with real-world potential—and a few sleepless nights.
Null