Dinevia

Dinevia — Menus Without Borders. A restaurant menu can be on your phone and still leave you unsure what to order. Dinevia turns a menu link or photograph into a readable menu in your language, with food preferences and approximate prices in your currency. No guest account is needed. Inspired by my experience in Moldovan restaurants, the mobile-first guest flow supports English, German, Spanish, Italian, French, Chinese and Japanese. It keeps dish descriptions, portions and original prices together, and retains the original so a guest can show the waiter exactly what they want. “For you” adds explained suggestions and ingredient conflicts to the same menu, without misleading match percentages. Meta Muse Spark performs photo recognition, translation and preference reasoning. The application validates structured outputs and numeric details, removes unrelated website content, and caches recognition and translations. Switching languages reuses the extracted menu. Currency conversions use published National Bank of Moldova rates and deterministic arithmetic. The restaurant vision is a prepared multilingual menu shared by QR code while retaining its visual identity. Image reconstruction is experimental; self-service host onboarding is not enabled in the public guest demo. Demo: a 2:34 edited walkthrough using actual local application screens and an owned fictional menu, with English narration and captions. The public GitHub includes local setup instructions. Deployment limitation: the website and demo are online, but hosted Meta inference currently returns 404 model_not_found on Render despite successful local calls. Please use the documented local setup with a Meta API key for functional evaluation. We do not claim the hosted translation is operational. Read every dish. Find your favorites. Feel at home at any table. Website: https://dinevia-1opu.onrender.com/