Logo Lanfrica

AnasKhaled1876/egypto_ai

Domain:

natural language processing

Record type:

software
Creator:
Ana
Host:
A culturally aware Flutter AI assistant for Egyptians and visitors exploring Egypt. egypto Ask smarter. Get better answers. A culturally aware AI assistant built for Egyptians and visitors exploring Egypt. egypto combines streaming AI chat with Egypt-aware guidance, web-assisted answers, voice input, response controls, guided discovery, and personal collections. The Flutter client supports Arabic and English from one feature-first codebase, with localization scaffolding for ten additional locales. The client builds with zero analyzer diagnostics and 57 passing unit, widget, and entity tests, gated in CI on every change. ## Product experience - **Streaming conversation:** cancellable Server-Sent Events deliver partial answers without freezing the interface. - **Text, voice, and web-assisted prompts:** users can choose how they ask and tune the response style without changing the backend protocol. - **Discovery and organization:** guided topics help start useful conversations, while collections keep important answers easy to revisit. - **Arabic and English first:** Arabic right-to-left layouts and English are supported today; ten additional ARB locales provide scaffolding for future reviewed translations. ## Architecture ```mermaid flowchart LR User["Android / iOS user"] --> App["Flutter client"] App -->|"Firebase ID token"| Auth["Firebase Authentication"] App -->|"HTTPS + cancellable SSE"| Api["Node.js / Express API Vercel Function"] Api -->|"streamed response"| App Api --> Redis["Managed Redis"] Api --> Db["Supabase PostgreSQL"] Api --> Models["LLM providers"] App --> Telemetry["Firebase Analytics Crashlytics / Messaging"] Cron["Vercel Cron"] --> Api ``` This repository contains the mobile client. The backend lives in a separate repository and is currently deployed on Vercel, with a configured maximum function duration and Vercel Cron for scheduled Explore work. A separately maintained Terraform and EC2 path is documented below as an alternative deployment design. The application uses a feature-first structure. Feature …