# Moment

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Tugdual de Reviers](https://cerebralvalley.ai/u/Tug)
- **GitHub:** https://github.com/Tugduoff/moment-launcher
- **Demo video:** https://youtu.be/UQFkIfuxYsw
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/5318

Moment is a calm Android home screen with one living companion. Instead of thirty icons, it shows a clock, the companion, at most four apps or actions for this moment, and search. At the uni restaurant at noon: “Pay with Izly” and directions. In the car: three big buttons. At the library: distracting apps step aside, and opening one means a 10-second wait; their notifications are held and summed up after. Late at night away from home: a ride home, share location, call a friend.

On-device rules decide when a context applies (time, place label, Bluetooth, calendar, visit patterns). Muse Spark decides what goes in it. When the companion notices “you come here most weekdays at noon”, it proposes a context on a card you accept, edit or dismiss: nothing changes without your OK.

You can simply talk to it, typed or by voice. It’s a Muse Spark agent with real tool calling: “add Spotify to Gym”, “remind me to call Léa at 7”, “I go to the gym Tuesday and Thursday evenings” edit contexts, set reminders, call favourites and remember what matters, with every change staged on one card. Snap your class timetable and Muse Spark vision proposes your week.

The companion is alive: Muse Image drew its body and edits it into outfit layers that stack instantly (hats, glasses, a ukulele), and the phone draws its face at 60 fps. It nods to your music, opens an umbrella in real rain, gets sleepy after midnight, sparks while charging, and gets dizzy if you shake the phone. Make it yours: face styles, body and leaf colours, a clock font apart from the app font, 12 living effects (neon, fireworks, sakura, code rain…), backgrounds it paints for each context or your own photo, and themes you save, tie to a context and share as a code. Morning and evening briefings, widgets per context, gestures, and a search that puts the best match right above your thumb make it a real everyday launcher. Everything sensitive (calendar, contacts, approximate location, steps) is opt-in on one “What I can see” page, and only short labels ever reach the model.

## More from Meta Global AI Developer Hackathon

- [anivns](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/5315)
- [Studio for One](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/5316)
- [PetPal](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/5317)
- [Backlot](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/5319)
- [Newton & Carter LLC](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/5320)
- [Accord](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/5321)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/5318. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt
