Manthana – Medical Domain Intelligence Platform

Manthana is a vertical medical domain intelligence platform composed of multiple FastAPI microservices orchestrated by a central AI router. It integrates imaging, signal processing, NLP, cheminformatics, and Ayurvedic intelligence into a single deployable stack.

Architecture

Frontend (User Facing)

• Manthana UI → frontend/manthana/ → port 3001
  → Built with Next.js + Stitch AI designs
  → The ONLY thing end users see

Backend Services

• AI Router → services/ai-router/ → port 8000
• ECG → services/ecg/ → port 8102
• Eye → services/eye/ → port 8103
• Cancer → services/cancer/ → port 8104
• Pathology → services/pathology/ → port 8105
• Brain → services/brain/ → port 8106
• Segmentation → services/segmentation/ → port 8107
• NLP → services/nlp/ → port 8108
• Drug → services/drug/ → port 8109
• Ayurveda → services/ayurveda/ → port 8110
• Imaging Utils → services/imaging-utils/ → port 8111
• Indexer → services/indexer/ → port 8112
• Auth → services/auth/ → port 8200

Internal Search Engine (Never Exposed)

• Perplexica → backend/search-engine/perplexica/ → port 3000
• SearxNG → search/searxng/ → port 4000
• Meilisearch → search/meilisearch/ → port 7700

Infrastructure

• Traefik → gateway/traefik/ → port 80/443
• Qdrant → storage/qdrant/ → port 6333
• n8n → orchestration/n8n/ → port 5678
• Grafana → observability/grafana/ → port 3002
• Prometheus → observability/prometheus/ → port 9090
• Loki → observability/loki/ → port 3100
• Ollama → ai/ollama/ → port 11434

All services share common utilities in services/shared (config, models, helpers) and follow a consistent JSON response envelope with request IDs and rate limiting.

Prerequisites

• Docker and Docker Compose
• For production: see DEPLOYMENT_CHECKLIST.md — especially MEILI_MASTER_KEY
• Python 3.10 (if running services outside Docker)
• Ollama installed with the required models (e.g. meditron, llama3.2)

Quick Start

cd medical-search-engine
cp .env.example .env
make build
make up
make health

Once running:

• Router health: http://localhost:8000/health
• Versioned API: http://localhost:8000/v1 (e.g. /v1/search, /v1/query)
• Router docs: http://localhost:8000/docs
• Individual service docs (examples):
  • ECG: http://localhost:8102/docs
  • NLP: http://localhost:8108/docs
  • Ayurveda: http://localhost:8110/docs

Development

Each service directory contains:

• main.py – FastAPI application
• requirements.txt – Python dependencies
• Dockerfile – container build
• .env.example – service-specific environment variables
• README.md – service documentation

You can run a service locally with:

cd services/ecg
cp .env.example .env
pip install -r requirements.txt
uvicorn main:app --reload --port 8102

Testing

Tests are placed under tests/ and can be executed with:

# Full suite (requires optional deps for drug/ecg/eye)
PYTHONPATH=. pytest tests

# Critical path tests only (no ML deps)
make test-fast
# or: PYTHONPATH=. pytest tests -m "not integration" -v --cov=...

# Exclude integration tests (rdkit, neurokit2, pydicom)
PYTHONPATH=. pytest tests -m "not integration"

Coverage target: 55%+ on api, orchestrator, plagiarism_service, services/shared, services/ai-router.

Disclaimer

All AI analyses provided by Manthana are for research and educational use only and are not a substitute for professional medical advice, diagnosis, or treatment. Always consult a qualified healthcare provider for clinical decisions.