camp helps people at work order meals without ever leaving their desk. It lives in the Mac notch and as a Live Activity on iPhone, uses your order history, preferences, and calendar to suggest what to eat, and groups coworkers who pick the same place so they can share the delivery fee.

Why

Everyone has dealt with long lines for coffee or lunch at work. camp makes ordering food at work as frictionless as possible: your order, on your time.

How it works

Recommendations start from your order history, dietary preferences, and calendar availability, so camp knows when you actually have time to eat and what you tend to like. If nothing fits, you can describe what you're craving in plain language and camp matches it against real menu items and prices instead of letting the model invent results.

When coworkers choose the same restaurant, camp batches their orders into one group so the delivery fee is split. Group membership, recommendations, and order state stay in sync across your laptop and phone, so the Mac notch panel and the iPhone Live Activity always show the latest state.

Features

• Personalized meal recommendations from order history, preferences, and calendar
• Natural-language food search grounded in real menus and prices
• Group ordering that batches coworkers’ orders to share delivery fees
• Mac notch panel built with AppKit
• Interactive iPhone Live Activities via ActivityKit
• Ramp integration for managing spend
• Laptop and phone stay in sync

How we built it

Native Mac and iPhone apps written in Swift: an AppKit panel for the Mac notch and ActivityKit-powered Live Activities on iPhone. A Python FastAPI backend connects the apps and stores profiles, groups, and orders. The recommender filters and scores menu items, then batches office orders to share delivery costs, adapting to user feedback over time. Natural-language search uses the OpenAI API to match cravings with menu options, and the recommendation path can also run without an LLM key.

My part was the recommendation and group-order backend: the restaurant catalog, ranking and batching logic, the API and database flows, and the LLM-powered restaurant search. Alex built the native Mac and iPhone experience, including the notch panel, the Live Activity, and the location, calendar, and Ramp sandbox integrations.

What's next

Integrating with real delivery partners through services like the DoorDash CLI, and pulling recommendations from your Beli rankings.

Awards

• 1st Place, Ramp: Save Time. Save Money at HackMIT 2026

Tech stack