AI Symptom Checker
Analyzes multi-symptom clusters across 15+ disease categories using a tri-model ensemble (Naive Bayes, SVM, Random Forest) hosted on FastAPI. Delivers clear medical explanations via Gemini 1.5 Flash.
HealthHub connects fragmented healthcare services into a single, cohesive companion. Powered by ensemble machine learning diagnostics, decentralized IPFS medical vaults, 24/7 AI triage, and real-time emergency dispatch.
Patients struggle across disconnected clinics, misplaced paper records, alarmist internet searches, and delayed emergency logistics. HealthHub builds a continuous, reassuring bridge.
Every feature on HealthHub is implemented in production code across Flutter, FastAPI, IPFS, Firebase, and OpenStreetMap.
Analyzes multi-symptom clusters across 15+ disease categories using a tri-model ensemble (Naive Bayes, SVM, Random Forest) hosted on FastAPI. Delivers clear medical explanations via Gemini 1.5 Flash.
Upload PDF prescriptions and lab scans directly to IPFS via Pinata cloud. Every upload calculates a SHA-256 integrity checksum. Generate 5-character share codes that grant doctors access for exactly 24 hours.
Filter top specialists by department (Cardiology, Dermatology, General Practice). View fees, credentials, and available slots. Seamlessly launch secure video appointments with pre-shared patient vault history.
Intelligent conversational triage built with Google Gemini 1.5 Flash. Includes Speech-to-Text and Text-to-Speech support, structured health literacy answering, and persistent chat sessions.
Camera-assisted pill identification, comprehensive OTC dosage guidance with age-specific recommendations and contraindication warnings, paired with the MediMaster interactive health quiz.
Locates nearby emergency hospitals using OpenStreetMap with distance color-coding. Enables one-tap ambulance bookings with live GPS coordinates, and provides a verified blood donation community registry.
HealthHub pairs deterministic ensemble machine learning for condition classification with Google Gemini 1.5 Flash for human-readable synthesis.
Implemented in disease_prediction_api_service.dart
and the FastAPI inference server.
Symptoms are queried dynamically by anatomical and clinical category (Respiratory, Digestive, Skin, Neurological).
User selections are transformed into standardized multi-hot indicator vectors matching the trained clinical symptom matrix.
Three distinct models evaluate the vector in parallel: Gaussian Naive Bayes, Support Vector Classifier, and Random Forest.
FastAPI reconciles model probability outputs to compute a calibrated confidence percentage and primary disease match.
Generates an empathetic 1-line layperson summary, practical non-medicinal precautions, and links directly to relevant doctor booking.
Select a clinical category and test the ensemble inference response:
Experience the real HealthHub mobile UI. Every layout, color, and interaction was built to reduce cognitive load during medical moments.
Guiding patients through early disease prediction with clear visual touchpoints.
Connecting patients with nearby verified specialists and available appointment slots.
Promoting preventative lifestyle recommendations tailored to user health history.
Role-separated entry for patients and doctors backed by Firebase Auth and Firestore.
Real-time doctor appointment check-in with synchronized patient history review.
Automated certificate issuance signed and timestamped upon verified blood donation.
We make zero unsupported claims. HealthHub implements content-addressed IPFS storage, SHA-256 integrity checksums, and strict 24-hour doctor share codes.
Traditional health platforms store records on centralized servers vulnerable to breaches. HealthHub stores encrypted document payloads on IPFS and allows patients to share time-bounded access directly with physicians.
Documents pinned to IPFS through Pinata cloud receive immutable Content Identifiers (CIDs) that cannot be altered or overwritten.
Every file upload generates a SHA-256 hash stored in Firestore, verifying that the document downloaded by a doctor matches the original file exactly.
Patients generate temporary alphanumeric codes (e.g. X7K9P) with strict 24-hour
expiration windows, preventing open-ended record exposure.
HealthHub indexes clinically validated OTC protocols directly from the in-app pharmacopeia dataset
(assets/json/medicine_recommendation.json).
Calculates real-time distances to nearest trauma centers with visual proximity tags (Green <1km, Yellow 1–3km, Orange >3km). Direct click-to-call routing.
Transmits exact device GPS latitude & longitude coordinates to paramedic fleet dispatch, showing real-time driver status, vehicle unit ID, and estimated arrival time.
Instant matching for urgent blood group units (A+, B+, O-, AB+). Verified donors receive official digital recognition certificates signed and timestamped in-app.
Download the HealthHub Android APK or inspect the full open-source Flutter repository. Built with Flutter, FastAPI, and Google Gemini.