Finalist
Asaply
Asaply is a lightweight checkout engine that uses GPT-5 to translate natural language into structured orders.
Core trick
GPT-5 maps a free-text order onto the merchant's menu schema and returns structured JSON, flagging uncertain items for verification before checkout.
Tags
Description
Asaply is a lightweight checkout engine that uses GPT-5 to translate natural language into structured orders. Customers describe what they want in plain English, and Asaply interprets the request, matches it to a merchant’s menu, verifies uncertainties, and generates a ready-to-process checkout plan. This makes ordering as easy as talking, while saving merchants the need to build custom chatbots or rigid menus.
README excerpt
# GPT Checkout (MVP) Minimal demo: Google Sign-In (Firebase) + GPT-5 planner API that turns natural language café orders into strict JSON. ## Setup 1. Copy `.env.local.example` to `.env.local` and fill values: - Get Firebase web config + enable Google Sign-In - Create a service account (JSON) and paste values into server envs - Add your `OPENAI_API_KEY` ## Firebase client notes This project initializes the Firebase Web SDK on the client. The client uses the following NEXT_PUBLIC environment variables (you can also rely on the fallback values already committed in the repo): - NEXT_PUBLIC_FIREBASE_API_KEY - NEXT_PUBLIC_FIREBASE_AUTH_DOMAIN - NEXT_PUBLIC_FIREBASE_PROJECT_ID - NEXT_PUBLIC_FIREBASE_STORAGE_BUCKET - NEXT_PUBLIC_FIREBASE_MESSAGING_SENDER_ID - NEXT_PUBLIC_FIREBASE_APP_ID - NEXT_PUBLIC_FIREBASE_MEASUREMENT_ID Analytics is initialized only in the browser (not during SSR) and will be a no-op on the server or when analytics is not supported. 2. Install & run ```bash npm i npm run dev Open http://localhost:3000 Flow Sign in with Google Enter: Large oat latte + chocolate croissant at 12:30 pickup Receive strict JSON plan ## After generating - Replace `.env.local.example` with a real `.env.local` and run: ```bash npm i npm run dev End of prompt. Create all files and confirm the app starts.