Lino Labs·Future apps·stamp launcher: livelauncher: offline Filed 2026-05-15
future app · harnesses Maps history idea AI-enhanced

stamp

Your Apple/Google Maps history, but only the places you went back to. stamp surfaces your real haunts — the bench, the bar, the corner — as tiny map-card kliplets.

"Places that earned a return."

The hook

Anyone who's ever opened Maps and thought "wait, I've been here before." stamp reads location history, finds the repeat-visit pattern, and turns each haunt into a klippable map tile.

The klip-feed

ReplacesThe Foursquare check-in reflex (RIP)
Klip payload1×1 map tile + Apple/Google Maps deep-link
Profile signalA personal map dotted with "places that earned a return"

AI layer

What it doesA clustering model resolves GPS noise into real "places" (50–100m bucket with semantic dedup) and ranks them by haunt-score (frequency × recency × dwell). An LLM writes the one-line caption for each haunt ("Tuesday-after-work bar").
InputRaw location history + venue database + visit timestamps
OutputResolved venue + haunt-score + one-line caption
Why with > withoutWithout AI, GPS coords just point at parking lots and intersections. With AI, the haunt is the bar, the bench, the bookstore — and it has a name.

Open questions

Status

idea  Filed 2026-05-15 from brainstorm session.

Lino Labs · future-apps · stamp