Ask Africare

AI health intelligence · Cloudflare Workers AI · LangChain agents

4-agent RAG pipeline · Router → Retriever → Analyst → Fact-CheckerLive

Hello! I'm Africare Intelligence — a 4-agent RAG system powered by Cloudflare Workers AI.

Pipeline: Router → Retriever (WHO DON + ReliefWeb + GDELT) → Analyst → Fact-Checker

I fetch live WHO outbreak data before answering — no more hallucinating case counts. Ask me about any active outbreak and I'll show you my sources.

05:20llama-3.1-8b-instruct

AI Architecture

LangChain

ReAct agent orchestration

Workers AI

Llama 3.1 8B · <50ms

MLflow

Experiment tracking

FastAPI

Python backend

Agent Pipeline

1User query → LangChain Router

2Tool selection (WHO API / Vector DB)

3Context retrieval + injection

4Llama 3.1 inference @ Cloudflare edge

5Structured response + citations

Data Sources

WHO GHODisease incidence

DHIS2National health systems

Africare DBClinic telemetry

World BankHealth financing

Production backend

Full Python FastAPI + LangChain + MLflow backend in /backend — Docker-ready for Kubernetes.