Title: Messenger AI Support + COD Orders for Algerian Stores
Description:
I build an AI sales assistant for Algerian clothing stores drowning in Messenger DMs.
Problem: same questions 24/7 - price? size? delivery to Oran? - missed 2am messages = lost COD sales.
What it does:
Replies instantly in Algerian Darja (Arabizi / Arabic / French mix) via Ollama local LLM
Product search from store catalog, hours/FAQ answers
Collects COD orders: name, phone (DZ validation), 58 wilayas, commune → saves confirmed order
Ready for Yalidine parcel create + tracking (demo mode now, live with client keys)
Owner dashboard: orders, inbox with AI/human takeover toggle, product price/stock edit Stack: Next.js 14 TS, Supabase (Postgres), Ollama llama3.1, Meta Messenger webhooks, Yalidine API. Status: working local MVP - tested end-to-end (product Q → full COD order confirmed). Next: live FB Page + multi-store SaaS.
1
14
Building a real-time, cross-platform client management system.
I’m currently wrapping up a multi-platform app designed to help businesses manage their client portfolios seamlessly. It includes a web dashboard for desktop work and a native mobile app for on-the-go access.
The Tech Stack:
Frontend: React JS & React Native
Database: Cloud Firestore (for instant, real-time data sync across all devices)
It’s built to keep team data perfectly synchronized in real time—no page refreshes, no delays.
The project is still under developement.
0
64
i am updating a mobile app for a client who want brand new design with improved performance
0
75
APFlow, is an AI-powered accounts payable (AP) automation app. You upload an invoice (PDF/image), a LangGraph agent pipeline extracts the data (vendor, amounts, line items), matches it against purchase orders, auto-books invoices that are low-amount and high-confidence, and routes the rest to a human for approve/reject — all with full audit trail
0
55