Hooligan

An iMessage-native planning agent for deciding what to do, where to go, and how to spend free time.

Small planning decisions make people bounce between maps, notes, recommendation apps, saved posts, and group chats just to answer what to do next.

Inside the experience

What someone can actually do

  1. iMessage conversation loop through Photon
  2. Lightweight memory for tried and liked places
  3. Short conversational planning responses
  4. Scheduled morning and midday reminders
  5. Future work on opt-in controls and safer location handling

How the idea took shape

A text-first agent in iMessage that asks short follow-ups, remembers lightweight local preferences, avoids repeated suggestions, and supports scheduled nudges.

The best surface for lightweight planning may be the conversation people already use—not another app they have to remember to open.
Explore how it was built +

I designed the conversational behavior and built the Photon integration, preference memory, recommendation loop, repeat avoidance, and scheduled nudges.

  • TypeScript
  • Node.js
  • Photon iMessage SDK
  • OpenAI API
  • Local JSON memory

Where it could go next

Ambient usefulness quickly becomes a privacy question; memory and proactive nudges need visible, reversible controls before they deserve user trust.

More work