Work
Makara: an interest-matched voice chat app
A real-time voice chat platform designed around store policies, with a server-side rule-based economy: NestJS, LiveKit and Flutter, one code base on three platforms.
- Client
- Kodzoft — kendi ürünü
- Services
- Custom software
Goal
A social app where verified, pseudonymous users have short voice chats with people suggested by shared interests. The product was designed from the start around the “random/anonymous chat” rules Apple and Google tightened in 2026: a match starts with a suggestion card, and identity verification and moderation are part of the product.
What we did
Product and policy research: competitor analysis, translating store rules (App Store 1.2, Google Play 18+ restriction, child safety standards) into the product flow.
Back end: REST + Socket.IO with NestJS, Prisma/PostgreSQL, Redis queues, identity with Firebase Auth, low-latency voice rooms with LiveKit.
Economy engine: coin and XP/level rules entirely on the server; every balance change in a single transaction with an idempotent ledger entry; AdMob server-side verification and daily caps for rewarded ads.
Mobile and web: Android, iOS and web with Flutter; matching, chat, friends, messaging, reporting and blocking flows.
Safety and moderation: report queue, automatic restrictions, published community guidelines and contact channels.
Highlights
All economy constants live in one configuration file; the client reads them from the API, so price and limit changes need no store update. Voice (LiveKit) and identity (Firebase) are isolated as replaceable layers.