Daply.Publishing, with AI in the editor.
At Daply I architected a multi-tenant content platform: workspaces with multiple members, websites published from prebuilt templates, a custom editor with AI tools for drafting and scheduling, and a media pipeline that transcodes uploads to m3u8 (HLS) for streaming. Node.js microservices handle load and access management, and the backend answers most API calls in 200–500 ms.
Highlights
- Backend tuned to 200–500 ms API responses
- Handed over with 2+ hours of documented knowledge-transfer sessions
The problem
Many publishers on one platform, each with their own members, content, media and access rules, and video that had to stream smoothly at every connection speed without a separate pipeline per tenant.
What I built
A platform rather than a site: shared services with per-tenant isolation, AI inside the editor instead of a bolt-on chatbot, and a backend tuned until most calls returned in 200–500 ms.
- 01
Multi-tenant CMS
Workspaces with multiple members, websites published from prebuilt templates, and complete isolation between tenants on shared infrastructure.
- 02
AI editor & auto-scheduler
A custom text editor with AI tools for creating articles, plus AI-assisted scheduling that publishes to WordPress, Telegram and webhooks.
- 03
HLS media pipeline
Uploads transcoded asynchronously to multi-bitrate m3u8 (HLS) for smooth streaming on any connection.
- 04
Microservices, payments & security
Node.js microservices for load handling and access management, Stripe subscriptions, rate limiting and CSRF protection; caching and query optimisation brought responses to 200–500 ms.