title: EdgeFirst AI
emoji: 🔬
colorFrom: indigo
colorTo: red
sdk: static
pinned: true
license: cc-by-nc-4.0
EdgeFirst AI
Spatial Perception at the Edge
Open-source libraries and microservices for AI-driven spatial perception on embedded devices.
About
Au-Zone Technologies builds EdgeFirst, a platform for deploying AI perception on edge devices. We support cameras, LiDAR, radar, and time-of-flight sensors for real-time object detection, segmentation, sensor fusion, and 3D spatial understanding.
Our stack spans four layers: Foundation (hardware abstraction and inference delegates), Zenoh (pub/sub microservices), GStreamer (media pipeline plugins), and ROS 2 integration. All released under Apache 2.0.
Model Zoo
Pre-trained models validated on real edge hardware with full-dataset accuracy metrics and detailed timing breakdowns per device.
Browse the EdgeFirst Model Zoo →
Introducing the EdgeFirst Model Zoo → — how the grid is built, and why every published number links to the validation session that produced it.
Supported hardware
EdgeFirst Studio
EdgeFirst Studio is our MLOps platform for the complete perception development lifecycle — dataset management, model training, INT8 quantization, on-target validation, and deployment. Free tier available.
Links
| Website | edgefirst.ai |
| Studio | edgefirst.studio |
| GitHub | github.com/EdgeFirstAI |
| Documentation | doc.edgefirst.ai |
| Model Zoo | EdgeFirst/Models |
| Company | Au-Zone Technologies |
Open-source libraries (Foundation, Zenoh, GStreamer, ROS 2 integration): Apache 2.0 · Model Zoo validation results & card data: CC BY-NC 4.0 · © 2026 Au-Zone Technologies
NXP®, i.MX, eIQ®, Neutron, and Ara240 are trademarks or products of NXP Semiconductors. Hailo is a trademark of Hailo Technologies Ltd. Jetson is a trademark of NVIDIA Corporation. All other trademarks are the property of their respective owners.