Production AI models, one line to deploy
Deployed with one call to load_model(). Two models ship free with SDK v0.2 today — presence-core and occupancy-home-v1. The rest of the catalog is in development: join a model's waitlist and we'll email you when it enters beta.
People Tracking
load_model("people-tracking-v3")
Multi-person detection, tracking, and counting from pure depth — privacy-safe by construction.
Gesture Recognition
load_model("gesture-core-v2")
Swipes, holds, and hand poses for touchless interfaces, robust to lighting.
Forklift Safety
load_model("forklift-safety")
Person-near-vehicle events with zone rules and severity scoring, agent-ready.
SLAM Assist
load_model("slam-vio-v1")
Depth constraints for visual-inertial odometry — drift-free mapping on AMRs and drones.
Fall Detection
load_model("fall-detect-v2")
Elder-care grade fall events from depth silhouettes — no cameras watching, no faces stored.
Zone Heatmaps
load_model("heatmap-analytics")
Dwell, flow, and congestion analytics streamed straight to the fleet dashboard.
Room Occupancy
load_model("occupancy-home-v1")
Room-scale people counting and pet rejection from depth CV — runs on a Raspberry Pi, no GPU.
Obstacle Segmentation
load_model("obstacle-seg-v1")
Ground-plane removal and obstacle clustering for navigation stacks.
Presence & Approach
load_model("presence-core")
Debounced presence and approach detection from distance ranging. Works on every device.
Models that know your hardware
Every marketplace model ships in device-specific builds — quantized for the exact sensor, resolution, and NPU. Licenses are per model, per year, unlimited devices within one product line.
- Continuous accuracy updates via OTA
- Version pinning for certified deployments
- Custom model training available on Enterprise
# pip install spatialai — runs today, simulator included from spatialai import Camera cam = Camera("spatial-home") cam.load_model("occupancy-home-v1") # depth CV, no GPU cam.load_model("presence-core") # models compose for f in cam.stream(): occ = f.models["occupancy"] print(occ["people"], occ["pets"])