Provat App: Bengali News, Editorial & Job Exam Ecosystem
A cross-platform Bengali editorial reader, competitive exam digest, and publishing ecosystem powered by React Native, BullMQ, and Vite.
About The Project
Provat is an end-to-end digital publishing, daily editorial digest, and career preparation ecosystem crafted specifically for Bengali-speaking candidates preparing for BCS, bank, and government job exams. The platform pairs a high-performance cross-platform mobile app with an advanced editorial management suite and a fault-tolerant Express 5 API server.
The mobile application delivers a tailored reading experience featuring categorized feeds (Daily Editorials, Job Circulars, Current Affairs, Translations, Viva Prep, Focus Writing, and Exam Notices), offline bookmarking, dynamic font scaling, gesture-driven multi-image lightboxes, and controlled link previews.
The administrative dashboard offers a full-featured editorial newsroom workflow equipped with a custom Tiptap v3 rich text editor, drag-and-drop image reordering, OpenGraph automated metadata scraping, content scheduling, and community moderation workflows with Role-Based Access Control (RBAC).
Engineered with a distributed backend, the system leverages BullMQ and Redis for precision post scheduling, automated worker recovery crons, live Prometheus metrics collection, NoSQL injection sanitizers, and an on-the-fly Sharp/Cloudinary WebP image compression pipeline.
The Challenge
Delivering complex Bengali typography and dense editorial content across diverse mobile screen sizes requires zero-layout-shift rendering, intelligent offline caching, and responsive feed pagination. Simultaneously, the platform needed to handle automated scheduled publishing, concurrent link metadata extraction without rate-limiting penalties, and strict content moderation in real time.
The Solution
Architected a unified three-tier ecosystem: a React Native 0.86 / Expo SDK 57 mobile client with NativeWind styling and Redux Toolkit caching; a Vite 7 & Tailwind CSS v4 admin dashboard with Radix UI primitives and Tiptap v3; and an Express 5 backend orchestrating BullMQ background job queues, Redis 24h link preview caching, and automated WebP asset optimization.
Core Features & System Modules
Every feature was purposefully architected for speed, reliability, and ease of operation.
Categorized Editorial & Exam Feeds
Curated feeds for Daily Editorials, Job Circulars, Current Affairs, Translation Practice, Viva Tips, and Focus Writing with cursor-based pagination.
Bengali Typography & Rich Text Engine
Native Bengali font rendering with custom Noto Sans Bengali typography, responsive text sizing, and high-fidelity HTML parsing.
BullMQ Distributed Publishing Scheduler
Automated background worker queue allowing editors to draft, schedule, and automatically publish articles down to the exact minute.
Throttled OpenGraph Scraper Queue
Concurrency-limited client queue and Redis-cached backend scraper providing instant, rich link preview cards without server overload.
Gesture-Driven Image Lightbox
Full-screen zoomable and pannable image gallery supporting multi-image grids with smooth Reanimated transitions.
Tiptap v3 Editorial Studio
Next-gen WYSIWYG editor supporting typography styles, KaTeX formulas, tables, syntax-highlighted code blocks, and drag-and-drop media grids.
Offline Storage & Bookmarking
Persistent offline storage powered by Redux Persist and AsyncStorage, enabling users to save crucial articles and read without internet connectivity.
Community Moderation & Reporting
Granular reporting system enabling users to flag inaccuracies, with dedicated dashboard queues for admins to resolve disputes swiftly.
Push Notifications & Alerts
Expo Push Notifications and Firebase Cloud Messaging (FCM) integration for real-time breaking news, exam circulars, and editorial drops.
Prometheus Metrics & Observability
Production-grade observability with Prometheus HTTP request histograms, counter metrics, Winston logging, and Swagger OpenAPI 3.0 docs.
Architecture & Technical Highlights
Dedicated Three-Tier Monorepo Separation
Segregated the client mobile app (`React Native / Expo SDK 57`), editorial back-office (`React 19 / Vite 7 / Tailwind CSS v4`), and API core (`Express 5 / TypeScript`) into decoupled micro-services for independent scalability.
Asynchronous Scheduling & Recovery Architecture
Implemented BullMQ with Redis to manage scheduled post publication queues, supplemented with automated fallback cron jobs to auto-recover delayed publications during server restarts.
Concurrency-Managed Link Preview Engine
Built a client-side fetch queue limiting concurrent metadata requests to 2 simultaneous threads alongside a 24-hour Redis server cache, eliminating rate limits and maximizing feed scroll frame rates.
Dynamic On-The-Fly WebP Asset Pipeline
Engineered a stream-based image processing pipeline leveraging Multer, Streamifier, and Sharp to automatically convert and compress all newsroom media assets into optimized WebP formats on Cloudinary.
Deep Defense Security & Rate Limiting
Integrated Helmet headers, strict recursive NoSQL object sanitization stripping malicious query operators, XSS filtering, windowed rate limiters, and complete Zod runtime schema validations.
Technologies & Frameworks Deployed
Have a Complex Project in Mind? Let's Build It Right.
We engineer bespoke software, scalable web architectures, and high-concurrency platforms that drive real business growth.