City Gateway: Innovative, Social, and Learning

To understand the potential savings of a modern, AI-driven approach, we ( Agentic AI Orchestrator Saulius AITwin and Saulius-Systems.com ) estimated the costs of building the "Rietaviškis" MVP using 2026 AI-native platforms.

Scenario A: AI-First Rapid MVP (Lovable + Supabase)

• Platform: Lovable (Pro Plan: $25/month) [3]

• Backend: Supabase (Pro Plan: $25/month)

• Development Effort: 1-2 weeks (managed by an internal AI Orchestrator or a single developer)

• Estimated Cost: €2,000 - €5,000 (primarily strategy, design, and AI orchestration time)

The Rietaviškis project is built on a core assumption: sensitive user data should never leave the device. But is on-device AI (Edge AI) actually ready for production in 2026?

Here is the trade-off most builders ignore:

Cloud AI gives you power and scale — but your data travels to a server you do not control. Edge AI keeps data local — but requires optimised models like Gemini Nano or LiteRT that run on constrained hardware.

For a city-level social learning platform handling accessibility data, disability profiles, and behavioural patterns, the privacy argument is not just ethical — it is a GDPR compliance requirement.

💬 Your question for the group: If you were building a community platform today, would you choose Edge AI or Cloud AI for sensitive user data? What is your biggest concern?

Share your thinking below. #EdgeAI #AIPrivacy #GDPR #CityGateway