moussdiop240-source/cahier-caisse-ai-showcase

Voice-driven customer credit ledger for Senegalese merchants (Wolof/French). Source private. # Cahier Caisse AI Most small shopkeepers across Senegal still track customer credit — who owes what, who paid back how much — in a paper notebook (a "cahier de caisse"). It works, but it's slow, error-prone, and impossible to search when a customer disputes a balance. Digitizing it usually fails because typing isn't how these merchants naturally record a sale — speaking is. Cahier Caisse AI replaces the paper notebook with voice. A merchant just speaks a sentence in Wolof, French, or the code-switched mix common in everyday Senegalese speech — "Moussa a pris à crédit deux mille francs" — and the app transcribes it, understands the intent (new credit, repayment, or balance summary), and updates the right customer's ledger automatically. No forms, no typing, no learning curve. ## Key capabilities - Voice-driven transaction entry in Wolof, French, and mixed Wolof/French speech - Automatic intent detection: new credit, repayment, or balance summary - Per-customer debt ledger, always up to date and searchable - Installable as a mobile app (PWA) for use directly from a phone - Mobile payment integration for subscription billing (PayDunya) ## Stack Next.js, TypeScript, Supabase, OpenAI Whisper --- Source code is private. Read-only access available to hiring teams and partners on request: moustaphaleye.diop@gmail.com