The spatial stack for location-aware AI
Modular and composable — use one layer or all three. Each pillar is a complete product; together they form a full spatial intelligence platform.
AI agents that understand their surroundings
Rich, AI-ready place data via MCP — stateless search, live position-aware context, or fully offline via ZephrTiles. Built on Overture Maps with AI-generated descriptions and entrances mapped to reality.
- MCP Lite — stateless place search, any AI system
- Egocentric MCP — live position-aware spatial context
- ZephrTiles — offline, on-device, no cloud dependency
Positioning that works indoors and out
GNSS with full 6DoF orientation delivers a consistent, sub-meter fix outdoors and through urban canyons, today. Visual Inertial Odometry (VIO) and Visual Positioning (VPS) for seamless indoor handoff are coming soon. Any Android, no extra hardware. ZephrFix for field use, or integrate via the Positioning SDK.
- Sub-meter GNSS accuracy, 10Hz update rate
- Full 6DoF localization
- Indoor VPS handoff (coming soon)
On-device intelligence, spatially grounded
Run local LLMs on mobile with full spatial context from ZephrTiles. Privacy-first, works offline, no cloud dependency. The conversational AI layer for local search and navigation.
- On-device LLM with real spatial context
- ZephrTiles for offline place grounding
- No data leaves the device
// GNSS corrections powered by
RX Networks TruePoint →What Zephr solves
Three products. Each one tackles a distinct layer of the location problem — from raw precision to spatial understanding to on-device AI.
your app knows where it is.
does it know what's around it?
Get API access, read the docs, and ship a location-aware AI feature this week. Questions? inquiry@zephr.xyz



