• 

  Front web page

• 

  Stats

• 

  Voice Chatbot

Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for ThalAssist

🩸 Project Story: EBloodConnect 🔍 Inspiration Thalassemia patients across India often struggle with regular blood transfusions, lack of donor availability, and minimal awareness. Witnessing this gap inspired us to build a tech-driven, AI-powered platform that brings together patients, donors, NGOs, and healthcare providers on a unified interface. Our goal: no Thalassemia patient should suffer due to a lack of timely blood support.

🛠️ How We Built It We developed the project using a modular stack:

Frontend: HTML, CSS, JavaScript, Tailwind (Vite)

Backend: Python (FastAPI), SQLite

AI Integration:

Voice-based Chatbot (using OpenAI + Web Speech API)

RAG pipeline for intelligent FAQs

Tool calling support for partner onboarding and demo requests

Donor Prediction: Past donor history analyzed via AI models for predictive availability

Real-time blood availability updates and camp notifications using secure APIs

Dashboard: Tracks donors helped, messages of gratitude, upcoming drives

Database: Donor, patient, request, and response logs securely stored

🤖 Key Features Voice-enabled AI Chatbot with dynamic tool-calling

Real-time donor matching + e-RaktKosh integration

Blood camp alerts and upcoming drives

Data privacy with secure storage & limited access

Gratitude wall: Donor appreciation, testimonials, and stats

🧠 What We Learned Building an AI-first healthcare interface from scratch

Integrating voice chatbot and RAG in real-time

Understanding data privacy regulations

Working with healthcare APIs and scaling for social impact

⚠️ Challenges We Faced Tailwind CSS and Vite setup issues

Ensuring accurate voice recognition

Handling async API calls with real-time UI rendering

Designing an interface that’s both tech-friendly and accessible

🧮 Future Scope Integrate with government health records

Build a mobile-first PWA

Deploy AI for blood group prediction from health data

Expand chatbot to regional languages

🧪 Math & Logic (Example) To calculate donor availability prediction:

Let

𝐷

donor history vector D=donor history vector

𝑇

time since last donation T=time since last donation

𝑃

probability of availability P=probability of availability

Then,

𝑃

𝜎 ( 𝑊 1 𝐷 + 𝑊 2 𝑇 + 𝑏 ) P=σ(W 1 ​ D+W 2 ​ T+b) Where 𝜎 σ is the sigmoid function

Built With

• css3
• flask
• github
• google-colab
• google-web-speech-api
• html5
• javascript
• langchain
• llm
• ollama
• python
• rag
• scikit-learn
• toolcalling
• vector-database
• webspeechapi