The problem
Arabic readers need a comfortable RTL experience, while translators need far more than a basic editor: assignments, chapter workflows, proofreading, summaries, moderation, and reader feedback.
Back to selected work
02Project mission dossier
Arabic publishing platform · 2026
Riwayat
A production Arabic webnovel platform with an RTL reader, structured translator publishing, administration, and context-aware AI tools across web and Android.
- Role
- Full-stack developer
- Status
- Deployed to production
Project snapshot
- Role
- Full-stack developer
- Status
- Deployed to production
- Platform
- Web and Android
- Deployment
- Vercel, Neon and Cloudflare
- Scope
- Reading, publishing, administration and contextual AI
- Repository
- moaadahtchaou/arabic-webnovel-platform
The solution
Riwayat combines a responsive reader, role-based publishing, secure administration, an Expo Android client, and server-side AI features that understand chapter context and respect spoiler boundaries.
Technical system
Verified components and boundaries found in the project source and documentation.
- 01RTL web client
React reader and publishing workspaces
- 02Android client
Expo app with SecureStore sessions
- 03Express API
Roles, CSRF, rotating refresh sessions
- 04Prisma + Neon
PostgreSQL publishing and context models
- 05Context pipeline
Translation, summaries and spoiler-safe Q&A
3role-based workspaces
Web + Androidshared platform
RTLArabic-first UX
What it does
Arabic-first responsive reading
Translator and admin workflows
AI translation and spoiler-aware Q&A
Shared secure API for web and Android
Product walkthrough02
Engineering decisions
Secure cross-domain sessions
The web client, API, and native client use different credential and domain boundaries.
- Decision
- Use rotating refresh sessions, HttpOnly cookies, CSRF protection, restricted CORS, rate limits, and role checks.
- Engineering effect
- Web and Android share one API without exposing server credentials to either client.
From server to serverless
The original Express runtime assumed a long-running server and local PostgreSQL.
- Decision
- Separate application export from local startup for Vercel and migrate relational data to Neon through Prisma.
- Engineering effect
- The same API model runs in production with managed PostgreSQL and custom domains.
Ownership map
Full-stack developer
- 01
Built the RTL React experience and Express API
- 02
Designed PostgreSQL models and authorization
- 03
Created AI, Android, deployment, DNS, and production reliability workflows
Outcome
The project evolved from a local prototype into a deployed multi-client product with a managed database, custom domains, secure cross-domain sessions, AI assistance, and production hardening.
Picnexa
Image processing infrastructure