ManyaFirmTech/manya-parse-ke

# Manya-Parse-KE 🇰🇪 **Open-source M-Pesa PDF statement parser with transparent financial health scoring.** Built by Manya Firm Technologies Ltd — Building Africa's Digital Future. --- ## The Problem Over **10 million Kenyans** are blacklisted or underscored by credit systems that use black-box algorithms. These systems penalize users for: - **Fuliza cycles** — even when caused by stock cycles, not bad behavior - **Off-net transfers** — sending money to Airtel, treated as "unstable cash flow" - **Non-financial metadata** — phone battery level, number of contacts, GPS location - **Ghost debts** — loans already repaid but still listed on CRB reports Under the **CBK's Draft Financial Consumer Protection Framework (2026)** and the **ODPC Data Protection Act 2019**, every Kenyan has the right to know how an automated system made a decision about them. Yet the tools to exercise that right don't exist. Until now. --- ## What Manya-Parse-KE Does Turns a messy M-Pesa PDF statement into a clean, structured JSON with: - ✅ **Transaction categorization** — Utilities, Savings, Loans, Suppliers, Income, Airtime, Betting, Government - ✅ **Manya Health Score** (300–850) — a fully transparent, cash-flow-centric alternative score - ✅ **Ghost debt detection** — finds CRB errors with receipt-level evidence - ✅ **Auto-generated dispute letters** — citing the Data Protection Act 2019 - ✅ **Suggested credit limit** — based on actual cash flow, not metadata proxies - ✅ **Plain English + Swahili explanations** for every score point --- ## Quick Start ```bash # Install npm install -g manya-parse-ke # Parse a statement manya-parse-ke my-mpesa-statement.pdf # Save full JSON output manya-parse-ke my-mpesa-statement.pdf --json --out result.json # Auto-generate CRB dispute letters manya-parse-ke my-mpesa-statement.pdf --dispute ``` --- ## Example Output ```json { "manya_health_score": 720, "score_grade": { "letter": "B", "label": "Very Good", "swahili": "Nzuri Sana" }, "insight …