Pangoni

Renting in Kenya runs on spreadsheets, WhatsApp threads and M-Pesa messages. Listings are scattered across a dozen sites, many of them stale or fake, and landlords reconcile rent by hand. Pangoni set out to fix both sides of the market: listings that seekers can trust, and a single system for landlords and agents to run their portfolios.

The problem

• Unreliable listings. The same unit appears on many sites at different prices, often long after it has been let.
• Manual rent collection. Paybill payments get matched to tenants by eye, receipts live on paper and arrears in notebooks.
• Compliance overhead. Invoicing, KRA tax filings and accounting exports happen separately, after the fact.

What I built

I designed and built the platform end to end, across web, mobile and backend:

• Landlord & agent hub. An Angular PWA for properties, units, tenants, leases, invoices, payments, maintenance, viewings and CRM, with analytics dashboards and role-based access.
• M-Pesa rent collection. Daraja STK Push plus Paybill/Till reconciliation that matches payments to invoices automatically and issues digital receipts.
• Tenant apps. Flutter apps so tenants can pay rent, see statements and raise maintenance requests from their phones.
• Verified listings pipeline. Listings from the major Kenyan property portals and social media are de-duplicated, location-matched, expired and verified before they go live.
• Natural-language search. Gemini turns queries like "2-bedroom in Kilimani under 80k with parking" into structured filters that run on Algolia.
• Tax & accounting. KRA eTIMS e-invoicing, plus exports to QuickBooks, Xero, Sage, Zoho, Google Sheets and Tally.

Under the hood

The backend is serverless Firebase: roughly 400 TypeScript Cloud Functions across 30+ business domains, Firestore with security rules that are covered by tests, and separate production and UAT environments. Rent escalations post to an append-only ledger inside transactions, so every tenant's history stays auditable. WhatsApp and email notifications, PDF and Word document generation, and a public API complete the platform.

The result

Pangoni is in production and growing. According to pangoni.io, landlords and agents use it to manage 400,000+ rental units, and it has tracked over KSh 900 million in rent across 60,000+ invoices.