AI-native eGov platform for Cameroon powered by MCP — DGI, CNPS, World Bank
# eGov Cameroun — AI-Native MCP Platform
An AI-native eGovernment platform for Cameroon built with MCP (Model Context Protocol).
Users interact with DGI and CNPS services through natural language in French or English.
## Architecture
```
React Frontend (MCP Client) → MCP Server (FastAPI) → Government APIs
↓
Claude Sonnet 4.6
```
Full details: docs/ARCHITECTURE.md
AI strategy: docs/AI_STRATEGY.md
## MCP Tools
| Tool | Description | API |
|------|-------------|-----|
| `verify_cnps_matricule` | CNPS registration check | cons.cnps.cm |
| `get_statistical_data` | Economic indicators | api.worldbank.org |
| `calculate_vat` | TVA declaration (19.25%) | Internal |
| `calculate_cnps_contributions` | Social contributions | Internal |
| `get_tax_deadlines` | DGI/CNPS deadlines | Internal |
## Quick Start
```bash
# 1. Clone
git clone && cd egov-cameroun
# 2. Configure
cp backend/.env.example backend/.env
# Edit backend/.env — add ANTHROPIC_API_KEY
cp frontend/.env.example frontend/.env.local
# Edit frontend/.env.local
# 3. Run with Docker
docker-compose up --build
# → Frontend:
localhost
# → Backend:
localhost
# → API docs:
localhost
```
## Local Development (without Docker)
### Backend
```bash
cd backend
python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env # add your ANTHROPIC_API_KEY
uvicorn app.main:app --reload
```
### Frontend
```bash
cd frontend
npm install
cp .env.example .env.local
npm run dev
```
## Running Tests
```bash
cd backend
pytest -v
```
## Test Strategy
### What is tested
**Unit tests** (`tests/test_tools.py`) — 14 tests covering the 3 pure-logic tools:
| Test class | What it validates |
|------------|-------------------|
| `TestCalculateVAT` | TVA rate (19.25%), deadline computation across month boundaries, input validation |
| `TestCNPSContributions` | Salary ceiling cap (750k FCFA), per-employee breakdown, invalid salary rejection |
| `TestTa …