Viseron

Viseron is a self-hosted, local-only Network Video Recorder (NVR) and AI computer vision platform built by Jesper Nilsson (GitHub: roflcoopter) and released under the MIT License. It runs entirely on your own hardware via Docker, keeping all video processing and data on-premises with no cloud dependency. The project is actively maintained, with the latest release (v3.5.3) published in April 2026.

What It Is

Viseron is an open-source NVR platform that combines traditional continuous video recording with AI-powered computer vision capabilities. It is designed for home, office, or any surveillance use case where privacy and local control matter. Users configure it through a config.yaml file edited via a built-in web interface, and extend its functionality through a modular component system.

Core AI Capabilities

Viseron's feature set is driven by pluggable components that can be mixed and matched:

• Object Detection – Real-time detection and tracking using models like YOLO and Darknet
• Motion Detection – Smart motion detection with configurable filters
• Face Recognition – Identify known and unknown faces
• Image Classification – Classify images using AI models
• License Plate Recognition – Read and track vehicle license plates
• Advanced Live View – Low-latency streaming via WebRTC

Hardware Acceleration Support

Viseron is built to take advantage of dedicated AI accelerators and GPUs, reducing CPU load and improving inference speed. Supported hardware includes:

• NVIDIA CUDA GPUs
• Google Coral EdgeTPU
• NVIDIA Jetson Nano
• NPU and TPU devices

This makes it practical to run demanding AI workloads on edge hardware without a powerful server.

Deployment Model

Viseron is distributed as a Docker container, making setup straightforward across Linux-based systems. Users spin up the container, configure it via the integrated YAML editor in the web UI, and select the components they need from the Component Explorer. The platform is camera-agnostic, supporting any RTSP-capable IP camera regardless of brand.

Update: v3.5.3 – Recording Video Controls Visibility

The latest release, v3.5.3 (published April 30, 2026), is named "Recording video controls visibility," indicating continued active development focused on the UI and recording experience. The repository on GitHub shows ongoing commits to the dev branch as recently as May 2026, with 3,143 stars and 384 forks, reflecting a healthy open-source community. The project also participates in Hacktoberfest and welcomes contributions via its issue tracker and discussion forums.