Team 75

UpNext

UpNext turns your upcoming calendar events into personalized shopping recommendations, helping you prepare ahead of time with real Amazon products and a clear explanation of why each item fits.

Check your own project

Community votes recognize the build and do not affect official judging.

Review the project

Start with the source code, then open the demo or video if available.

View GitHub repository
Project description
UpNext is a proactive personal shopping agent that helps you prepare for what’s coming next. Connect Google Calendar, and UpNext uses Airbyte to sync upcoming events and their descriptions. A local Qwen model running through Ollama plans Amazon searches, reviews retrieved product information, and recommends relevant items with links, observed prices, ratings, and explanations. For example, a Thanksgiving potluck requesting a turkey hat and a dish can inform shopping ideas, while a moving-day event prompts research into moving supplies. Users can ask for additional recommendations through an event-specific chat. Research runs after calendar syncs or on a configurable daily schedule, skipping events that already have recommendations. The Home page brings purchase options together, while the calendar shows pending research, pending purchases, and purchased items. Users can mark products purchased or dismiss them, with decisions and explicit preferences saved per account. Each workflow retains an audit trail of calendar context, model exchanges, retrieved sources, timing, and results. Amplitude and Statsig SDKs support telemetry and feature controls in local mode. Purchases remain user-controlled: UpNext recommends and links to products without checking out.
Tools used