Foot‑Traffic AI That Counts Human Patterns, not Personal Data
Runs on a Raspberry Pi using YOLOv8 + DeepSORT, tap into your existing IP cameras, or we’ll recommend affordable options. We install, calibrate, and keep your data private, frames stay on‑site, only JSON outputs leave the device.
See Pricing & BookHow It Works
Step 1: Use Your Existing Cameras, or Choose from Our List
Many venues already have IP cameras. We’ll integrate with those. Need one? Here are compatible options you can order:
- ANNKE C300 PoE turret – $29.99
- Wyze Cam v4 Wi‑Fi 2.5 K – $51.96
- Raspberry Pi HQ Camera – $50.00
Step 2: Plug In. We Calibrate.
We visit your venue, install a Raspberry Pi (4 or 5), and calibrate the camera view so YOLOv8 + DeepSORT can accurately count people whether they walk by or inside.
Step 3: Edge Processing + Privacy First
All video frames are processed locally on the Pi. No personal data is stored, just anonymized counts exported as JSON. YOLOv8 handles detection; DeepSORT maintains per‑person tracking across frames.
Why It Works for Your Venue
- No inference to the cloud → faster, more private.
- Avoids blurry counting methods, tracks real human motion for accuracy.
- Fully automatic and maintenance-light once calibrated.
- No personal identifiers, just counts (JSON). GDPR/CCPA compliant.
Pricing
Included in your plan. See pricing.
Notes from the Field
YOLOv8 combined with DeepSORT provides object detection and identity tracking, key for accurate tracking even in blur or overlapping scenarios.
Running YOLOv8 on a Raspberry Pi keeps processing local and private.