← All posts

2025-09-26

Building Floai Studio

Next.jsWebSocketAIPerformance

Context

Floai Studio (evolved from earlier CreatorFlow work) started as a reaction to fragmented creator tooling: one tab for drafts, another for analytics, and yet another for scheduling. The goal was a single surface with real-time feedback when something went live or failed.

Architecture notes

The app shell is Next.js with a clear split between marketing pages and the authenticated product. Real-time updates use WebSockets so the composer and dashboard stay in sync without polling everything.

For AI-assisted captions, the integration is intentionally narrow: suggest variants, let the human edit, then post. That keeps latency predictable and avoids “magic” failures in the critical path.

Performance

We leaned on caching for read-heavy dashboards and kept payloads small on mobile. The biggest win was treating mobile as first-class in QA, not an afterthought.

Takeaway

Ship the smallest AI surface that saves real time, then expand. Users preferred reliability over flashy demos.