How kakumanu-harshitha/medintel-ai works
AI Health Assistant is a React/FastAPI health companion: users authenticate, submit text or voice and optional images/reports, receive informational clinical analysis, and can review saved history and reports. The backend harmonizes multimodal inputs, routes clinical questions through intent/RAG and LLM reasoning, and stores assessments and conversation history. Separate branches cover profile/authentication, report generation, and administrator monitoring. Edges reflect the sampled entry points and explicit README capabilities; several feature routers and integrations were not sampled, so their internal wiring is left out.
Main entry points: main.py FastAPI app · query_service.py Multimodal endpoint · llm_service.py Clinical reasoning · dashboard_service.py Owner monitoring
Components 23
Patient experience
- React routes
frontend_react/src/App.jsx - Assessment chat
frontend_react/src/pages/Chat.jsx - API client
frontend_react/src/services/api.js - Reports page
frontend_react/src/pages/Reports.jsx
Identity and profiles
- Authentication
backend/app/auth/logic/auth.py - Profile API
backend/app/api/profile_router.py - PostgreSQL
Clinical assessment
- FastAPI app
backend/app/main.py - Multimodal endpoint
backend/app/api/query_service.py - Clinical reasoning
backend/app/services/llm_service.py - Intent and RAG routing
backend/app/api/rag_router.py - Speech processing
backend/app/services/speech_service.py - Report extraction
backend/app/services/report_processor.py - Groq LLM
Knowledge and memory
- Medical retrieval
backend/app/rag/rag_service.py - Pinecone vectors
- Medical data ingestion
backend/medical_ingester.py - MongoDB history
backend/app/database/mongo_memory.py
Reports and oversight
- Report API
backend/app/api/report_router.py - Feedback API
backend/app/api/feedback_router.py - Owner monitoring
backend/app/admin/dashboard_service.py
Other components
- Patient
- Medical specialists
Connections 27
- Patient → React routes: uses
- React routes → Assessment chat: routes
- Assessment chat → API client: submits query
- API client → FastAPI app: API requests
- FastAPI app → Authentication: mounts routes
- FastAPI app → Profile API: mounts routes
- FastAPI app → Multimodal endpoint: mounts routes
- FastAPI app → Report API: mounts routes
- FastAPI app → Feedback API: mounts routes
- FastAPI app → Owner monitoring: mounts routes
- Authentication → PostgreSQL: reads and writes
- Profile API → PostgreSQL: reads and writes
- Multimodal endpoint → Speech processing: transcribes audio
- Multimodal endpoint → Report extraction: extracts report
- Multimodal endpoint → Clinical reasoning: requests analysis
- Clinical reasoning → Intent and RAG routing: routes clinical intent
- Intent and RAG routing → Medical retrieval: selects retrieval
- Medical retrieval → Pinecone vectors: searches vectors
- Clinical reasoning → Groq LLM: requests generation
- Multimodal endpoint → MongoDB history: stores conversation
- Medical data ingestion → Pinecone vectors: indexes medical sources
- Report API → MongoDB history: reads saved reports
- Reports page → API client: requests reports
- Feedback API → PostgreSQL: records feedback
- Owner monitoring → PostgreSQL: reads audit and profile data
- Owner monitoring → MongoDB history: reads history and reports
- Clinical reasoning → Medical specialists: flags escalation