A custom editorial platform with content import, three language versions, a visual builder, media library and a fast, SEO-ready portal.
Content arrives from an external platform, undergoes editorial review, is translated into Russian, Kazakh and English, receives media and a publication date, and then needs to appear on the website quickly. Manual transfers, disconnected translations and developer dependency slowed this cycle. We separated the product into an API, an editorial panel and an independent public frontend.
Articles, images, translations, menus and settings are managed separately, leading to migration errors.
We brought Article, Translation, Category, Media, Menu and Settings together in a shared model and a single working interface.
Editors cannot see which translations are ready, published or scheduled.
We made RU/KZ/EN independent entities with their own status, date, SEO fields and readiness controls.
Changing the homepage, section or block order requires code changes.
We created a drag-and-drop builder with 20 block types, draft and preview modes, and independent layout publishing.
Content from Insider has to be transferred, categorised and checked manually.
We added full and partial synchronisation, scheduling, category mapping and mandatory-language rules.
The architecture was verified using the supplied documentation, application structure and genuine product screens.
The article list displays statuses, categories, dates and language versions. Synchronization turns technical rules into understandable settings, and the visual builder lets editors assemble home, category and article pages without developer assistance. After publication, the backend invalidates only related cache entries, updating the site without a full rebuild.
Status, date, author, type, section, media and metrics
RU/KZ/EN, custom slug, SEO and publication time
Sections, columns, blocks, order, draft and version
MIME, dimensions, file size, alt, caption and Cloudinary URL
Source, result, errors and synchronization time
Roles, revocable sessions and change history
Tiptap for text, media and structured content
dnd-kit for sections, columns and 20 block types
Scheduled publishing and separate translation statuses
Views, bookmarks, comments and reactions are synced back to the CMS
The public frontend generates dynamic metadata, canonical URLs, language alternate links, Article and Breadcrumb JSON-LD, a standard sitemap, a Google News sitemap and robots.txt. SSR/ISR provide ready-to-read HTML for search engines, while targeted revalidation updates only changed content. SEO becomes part of the publishing pipeline rather than a manual editorial task.
The administration panel screenshots use demonstration data but were captured from the working interface. Scale is stated based on the current repository and is not substituted with traffic metrics.
At a width of 375 px, the public website preserves its sections, readable typography, illustrations and links between articles. Content remains server-rendered, and next-intl maintains the structure when switching between RU/KZ/EN.
The API covers authentication, content, media, synchronisation, layout and public data.
Three TypeScript applications separated by responsibility.
JWT, bcrypt, Zod, Helmet, CORS allowlist and separate rate limits.
Drafting, preview, scheduling, publishing and targeted cache invalidation.
Docker, Prisma migrations, health-check, Winston and error handling.
The REST API is ready for a mobile application, bot or additional storefront.
The business value lies in fewer manual transfers, less dependence on developers and unified control over publishing in three languages.
We will analyse the audiences, content, integrations and search demand, then design the structure and technical architecture before visual design begins.
Discuss the Project ↗Back to Case Studies