What problem does it solve? Local Flask apps with AI API backends fail when exposed to mobile or remote users because common tunnels drop long-running requests, CORS breaks cross-origin fetches, and free-tier AI APIs hit quota limits or return truncated JSON. ## Core Features & Use Cases - Single-Origin Flask Architecture: Serve frontend static files from the Flask backend to eliminate CORS issues and cross-tunnel fetch failures. - Tunnel Selection & AI API Handling: Use cloudflared for long-running requests, probe Gemini models for quota status, and recover truncated JSON responses with bracket-state parsing. - Use Case: Build a mobile-accessible video analysis dashboard where users upload videos from their phone, Gemini analyzes the footage, DeepSeek generates coaching feedback, and SQLite stores session history — all reachable through a cloudflared tunnel URL. ## Quick Start Ask the AI to build a Flask app that serves a frontend dashboard, integrates the Gemini API with model quota probing, and exposes it through a cloudflared tunnel for mobile access.