Speedykom/ZHSF-AI-claims-reconciliation-final

ZHSF AI Claims Reconciliation automates healthcare claims reconciliation processes for the Zanzibar Health Sector Fund using AI-powered workflows # ZHSF AI Claims Reconciliation This repository contains a local, Docker‑Compose based stack to prototype and develop the ZHSF AI claims reconciliation workflow. It combines workflow automation (n8n), data services (Supabase/Postgres), vector search (Qdrant), OCR, and a frontend for review. --- ## 1) What this project does The goal is to automatically reconcile healthcare claims by: - ingesting claim documents (PDF/images), - extracting text with OCR, - normalizing the data, - matching claims against reference records, - writing reconciliation results to a database, - and providing a UI for review and audit. --- ## 2) Quick start (local) ### Prerequisites - Docker + Docker Compose - Node.js (only needed if you want to run the frontend locally outside Docker) ### Step‑by‑step 1. **Configure environment** - Copy `.env.template` to `.env` and update values as needed. 2. **Start the stack** - Run: `make start` 3. **Verify services** - n8n: localhost - Supabase Studio: localhost - Qdrant: localhost 4. **Stop the stack** - Run: `make stop` --- ## 3) Services & Ports | Service | Purpose | Port | |---------|---------|------| | n8n | Workflow automation & orchestration | 5678 | | n8n-postgres | n8n internal database | 5432 | | Supabase Studio | DB management UI | 3000 | | Supabase (Postgres/API/Auth/Storage) | Core data services | internal | | Qdrant | Vector search / similarity | 6333 | | PaddleOCR | Document text extraction | internal | | Frontend | Review UI | 3000 (if run locally) | --- ## 4) Repository map (what to look at first) - `docker-compose.yml` → all services and networking. - `n8n/` → workflows, credentials template, and reference rules. - `supabase/` → database development assets and local setup. - `qdrant/` → vector DB configuration. - `paddleocr/` → OCR service. - `frontend/` → UI for reviewing reconciliation results. - `init-test-db/` → SQL seed data for local testing. - `training-onboarding-zhsf.md` → full …