NEPI Docker Container

NEPI runs on NVIDIA Jetson platforms (Nano, Xavier NX, Orin Nano, Orin NX, AGX Orin) and x86 systems with Docker support. All Numurus S2X hardware ships with NEPI pre-installed.

Yes. NEPI is open source under the BSD-3 license. The full source code is available on GitHub. Educational, professional, and enterprise license tiers are available for teams that need commercial support. See pricing for details.

No. NEPI's browser-based interface and low-code automation tools let you configure hardware, deploy AI models, and build workflows without writing ROS code. If your team does work with ROS, NEPI is fully compatible with standard ROS 2 nodes and packages.

NEPI supports deploying models trained in common frameworks including TensorRT, ONNX, and PyTorch. The platform manages how models connect to live sensor data, share outputs, and trigger downstream automation.

NEPI includes plug-and-play drivers for a range of cameras, LiDAR sensors, IMUs, GPS receivers, sonars, and actuators. If your hardware is not natively supported, the driver framework lets you write custom drivers that integrate with the rest of the platform. Check the documentation for the current supported hardware list.

The Docker container is a pre-built, pre-configured version of NEPI that runs on any supported platform with a single command. Building from source gives you full control over the build process and the ability to modify the platform itself. Both approaches give you the same NEPI platform.

NEPI (Numurus Edge Platform Interface) is a plug-and-play, low-code edge AI and automation platform built on ROS and ROS 2. It provides the core software every smart system needs to connect hardware, process data, run AI models, automate actions, and manage data at the edge. NEPI is often described as "Windows for smart systems" - just like Windows standardized how computers interact with hardware, software, and users, NEPI standardizes how smart systems integrate sensors, AI, automation, and data so teams don't have to rebuild the same foundation for every project.

NEPI is used by startups and early-stage teams building robotics, sensing, or automation products; researchers who need reliable, repeatable smart data collection systems; educators and student teams learning applied AI and automation; and engineering teams deploying AI-enabled systems in real-world environments. If your team needs to connect sensors, run AI at the edge, automate actions, and turn raw data into outcomes, NEPI is likely a strong fit.

NEPI eliminates the need to repeatedly build and maintain: hardware and sensor integration software, ROS and ROS 2 system plumbing, AI model deployment and orchestration pipelines, automation and event-driven logic, and data collection, synchronization, and storage systems. This dramatically shortens development timelines, helping teams move from concept to field-ready prototype in days or weeks instead of months.

No. NEPI is designed to make advanced edge AI accessible beyond traditional software teams. Using NEPI, users can bring up cameras, sensors, and devices quickly, run AI models without building custom infrastructure, and configure automation and actions using low-code tools. Advanced teams can still extend NEPI deeply when needed.

NEPI is built on top of ROS and ROS 2. Rather than replacing ROS, NEPI extends it by providing standardized drivers, AI management, automation tools, data handling, and user interfaces that many teams would otherwise need to engineer themselves. NEPI accelerates ROS-based development while remaining fully compatible with the broader ROS and ROS 2 ecosystems.

Most users can install NEPI, connect hardware, and begin running AI models within a single day. Documentation, tutorials, and example workflows are available to support fast onboarding.

Teams commonly use NEPI to run computer vision and perception at the edge, detect conditions and trigger automated responses in real time, collect high-quality synchronized smart data, and rapidly prototype and deploy AI-enabled systems. NEPI is intentionally industry-agnostic and supports science, education, and commercial applications.