# Introducing Detecto, on-premises AI video detection

_October 17, 2026 · LNX Inc._

Detecto runs AI detection on your existing IP cameras from one on-site GPU server. Free for schools and non-profits; LNX can deploy it.

LNX Inc. has built Detecto, a real-time video surveillance system that adds AI detection to the IP cameras a site already has. It runs on a single server on the premises, and video does not leave the site. Schools, colleges, universities, and non-profit organisations can use it free of charge, and LNX can deploy it for any organisation that wants help.

## What Detecto does

Detecto connects to existing IP cameras. It finds ONVIF cameras on the network automatically and can scan IP ranges for RTSP streams, so most sites do not need new cameras to start.

Live video streams to a browser dashboard over WebRTC or MSE. On the same server, YOLO models run on an NVIDIA GPU with TensorRT and look for:

- **People, vehicles, and animals.**
- **Weapons:** guns and knives.
- **Fire.**
- **Behaviour from pose analysis:** falls, fighting, arms raised, and crouching.
- **Motion.**

Detections are drawn as overlays on the live video. Each event is recorded with its history, and the event list can be filtered by severity. Every camera can have its own motion masks and zones, so a busy street at the edge of the frame does not raise alerts all day.

## Automations

Detecto includes rules in an "if this, then that" style. A rule has a trigger, optional conditions, and one or more actions.

- **Triggers** include a gun detected, a person detected, a camera going offline, and no motion for a period.
- **Conditions** narrow a trigger by time window, camera, zone, duration, confidence, and count.
- **Actions** include a browser notification, a webhook, a Slack message, an email, and recording a clip.

Templates for common rules are included, so a new rule can start from a template rather than a blank form. For example, a rule can say: when a person is detected in a chosen zone outside opening hours, email the facilities manager.

## Where it runs

Detecto runs fully on-premises on one Ubuntu server with an NVIDIA RTX GPU. Cameras, video, detections, and event history stay on that server. Nothing is sent to a cloud service unless you configure an action, such as a webhook or an email, to send it.

## What it is not

Detecto is an aid to the people who watch the cameras, not a replacement for them. Models miss things and flag things that are not there. Detecto helps staff notice events sooner and find them faster afterwards. Decisions about what an alert means, and what to do about it, stay with people.

## Licensing

- **Free for education and non-profits.** Schools, colleges, universities, non-profit and charitable organisations, public research institutions, and government bodies can use Detecto at no cost under the [PolyForm Noncommercial License 1.0.0](https://polyformproject.org/licenses/noncommercial/1.0.0/).
- **Commercial license for businesses.** Companies using Detecto in a commercial setting need a commercial license from LNX.

## How LNX can help

Software is one part of a camera system. LNX can deploy Detecto end to end:

- A site survey of your cameras, network, and coverage.
- Camera integration, including the cameras you already have.
- Supply and installation of the GPU server.
- Configuring detection, zones, and alert rules for how your site actually runs.
- Training the staff who will use the dashboard and respond to alerts.
- Ongoing support.

This builds on our [low-voltage and camera work](/services/low-voltage/) and our [AI & Automation](/services/ai-automation/) practice.

If you are a school or a [non-profit](/industries/nonprofit/), the software is free, and we can help you deploy it. If you are a business, we can quote the license and the installation together. Either way, tell us about your site on the [contact page](/contact/).