How Presque Built a Real-Time Spatial-Temporal Index for Memory Matching

Developers behind Presque, an iOS app that connects people who shared the same location at the same time, have detailed the engineering challenges behind its core matching system. Each user memory stores coordinates, a timestamp, a radius, and a personal time window, requiring matches to satisfy both spatial and temporal conditions simultaneously. While geospatial indexing handles location queries efficiently, time matching is complex because each memory carries its own variable window rather than a global fixed interval. A consent layer adds further complexity, requiring that visibility changes — including deletions and privacy toggles — propagate to the index instantly rather than through batch updates. The system also routes near-miss candidates, those just outside the spatial or time bounds, to a separate inbox rather than discarding them outright.