Technical Specifications

• Genie URL: https://gcsurge-docs.nxgen.cloud/gcsurge/docs/setup-deployment/technical-specifications

High-level engineering reference: how each part of GC Surge works, its limitations, and what will not work. Implementation details (tech stack, tuning parameters) are out of scope.

Onboarding — Adding Devices

Overview

Onboarding connects a customer’s cameras or alarm sources to GC Surge so their alarms flow into the Alarm Center. Every device belongs to a site; a site can hold many devices. Onboarding is reached from Configuration → Add, which offers four entry points:

A device is always attached to a site. Typing an existing site’s name adds the camera to that site instead of creating a new one.

Connection methods

GC Surge supports four ways for a camera’s alarms to reach the cloud. The guided wizard selects one automatically; the manual form auto-detects it from the reachability check and the camera brand.

Supported camera brands

A camera brand is supported when GC Surge can automatically inject the alarm configuration into the device — no manual device-side setup. The baseline supported brands are:

• Hikvision
• Dahua
• Axis
• NxWitness

The live list in the platform is authoritative — it may include additional brands beyond this baseline. Always treat the platform’s supported-brands dropdown as the current reference.

A camera not on the supported list is not blocked. It has two fallback paths:

• The camera can send its own alarms (API / FTP / SMTP). GC Surge generates the endpoint or address; the user pastes it into the camera’s settings.
• The camera only streams video. A GC Edge unit on site watches the stream and produces the alarms.

Guided wizard — decision logic

The wizard collects the minimum it needs and routes to the right method across three steps:

1. Intent — new site, add a camera to an existing site, or bulk add.
2. Camera type — a brand from the list (auto-configured); my camera can send alarms (API/FTP/SMTP) → manual path; use GC Edge → edge path; I’m not sure → continue and detect.
3. Access (only when not already decided) — GC Surge can reach it directly → Public; private network only → Private (local agent); not sure → Private (safest default).

Single device — fields collected

Bulk import — what is and is not supported

The spreadsheet importer supports: multiple sites and cameras in a single upload; all three connection types (public, private, edge); per-row validation before commit; a downloadable spreadsheet of failed rows; and a dedicated private device configuration step after import.

The manual protocols (API/SMTP/FTP) path is not supported in bulk import — that flow is per-device and interactive (the user copies connection details into each device individually).

NOVA99x — False Alarm Filter

NOVA99x is a noise-suppression module that filters up to 99% of false alarms, allowing security operations to scale without increasing headcount. It processes motion-triggered events in real time and classifies each event as a true alarm or a false alarm.

What it detects

NOVA99x uses the pre, current, and post images from each event (three frames) to detect motion and identify persons and vehicles — including personal cars, buses, trucks, and vans. It works with both optical and thermal sensors.

What it does NOT detect

• Animals
• Distant objects (small persons or vehicles far from the camera)
• Fast-moving cars
• Industrial vehicles and their parts

Camera compatibility

NOVA99x operates on event images rather than on specific camera streams — it is not tied to a particular set of camera models.

Limitations

• Cameras or alarm sources that send only a single image per event will have a very low true detection rate. NOVA99x is optimized for three-frame events.
• Video clips have lower throughput than three-frame events.
• Point cameras at the area of interest (entrance, clear field of vision). Avoid vegetation, support structures, or other obstructions that block the view.
• Detection accuracy decreases with distance and with lower illumination or contrast — night-vision cameras, thermal cameras, and low-light night scenes all reduce detection performance.

How it works (cloud)

NOVA99x processes the three event images in five steps:

1. Analyzes motion and identifies motion zones.
2. Analyzes image content and detects persons and vehicles.
3. Validates detection movement within the regions of interest and identifies stationary artifacts.
4. Filters out very fast and stationary objects, while keeping persons and moving vehicles.
5. Classifies the event based on the detections, the motion zones, and the configured parameters.

How it works on GC Edge (on-site)

On-site, NOVA99x analyzes the live feed and maintains a motion model. Once the motion model is triggered, the detection model activates to track objects and validate their movement.

GC Edge

GC Edge is the on-premise software appliance developed by NXGEN Technology AG that runs on a Windows PC at the customer site. It ingests video streams from IP cameras, runs AI motion detection locally, and publishes alarm events to GC Surge. All video processing happens on-device — no video is sent to the cloud.

Minimum hardware requirements

How it works

Pairing — During initial setup, the operator enters the site key generated from the GC Surge portal into the GC Edge Manager application. GC Edge uses this key to authenticate with the platform, retrieve the list of devices registered to the site, and connect to each one automatically. No re-pairing is needed on restart.

Video ingestion — Each camera stream is handled independently. GC Edge connects directly to the camera over the local network, pulls the live video feed, and processes it entirely on the local machine. Hardware acceleration is used where available to minimize CPU usage.

AI detection — Video frames are analyzed locally by an AI model that detects and tracks people and vehicles in real time. On static scenes with no movement, AI analysis is suspended automatically and resumes the moment activity is detected.

Monitoring — The GC Edge Manager provides a live dashboard showing the status of every connected camera — whether it is active, the current detection state, and a live thumbnail view.

Connectivity — GC Edge requires a stable internet connection to deliver alarms to GC Surge and to allow the live feed to be fetched remotely.

Limitations

• A camera must be reachable on the local network; cameras on other networks or VLANs are out of reach.
• Detection quality depends on camera placement and scene conditions (same caveats as NOVA99x above).
• Capacity is bounded by the host PC — camera count scales with available CPU, memory, and GPU.
• A stable internet link is required for alarm delivery; while offline, alarms cannot be delivered to the platform.

Genie Discovery Agent (GDA)

The GDA is the Private (local agent) connection path. It configures cameras on a private network that cannot be reached directly from the cloud: the agent scans the site’s local network and injects the alarm configuration into supported cameras it finds. Unsupported or unreachable cameras are skipped.

The GDA comes in two forms — a desktop agent (Windows) and a mobile app — that perform the same job.

Option A — Desktop Agent (Windows)

Use this when you are already connected to the same network as the cameras, or can remote into a machine that is. Install the GDA on any Windows PC on that network. Once running, the agent automatically detects which cameras belong to the site and configures them without further input. The portal status updates to Configured once the agent finishes. No on-site visit is needed if you already have remote access to the client’s network.

Option B — Mobile App

Use this when you do not have network access to the client site and need to send someone there in person. When cameras are added for a site with no local agent, the system generates a site key and QR code. Share this with the on-site person. They:

1. Install the GDA mobile app on their device.
2. Scan the QR code or enter the site key manually to link the app to the site.
3. Connect their device to the same local network as the cameras.
4. The app finds all cameras for the site and configures each reachable one automatically.
5. The portal status updates to Configured once complete.

The mobile app has no login — access is controlled through the site key. Only share the QR code with trusted personnel.

Supported camera brands

• Hikvision
• Dahua
• Axis
• NxWitness

Limitations

• The PC or mobile device must be on the same local network as the cameras during setup. Neither can reach cameras across different networks or VLANs.
• The mobile app has no login — access is controlled through the site key only. Share the QR code with trusted personnel only.
• If a camera is offline or unreachable at setup time, it will not be configured automatically. It can be retried once back online.
• Each site key is tied to a single site. If cameras span multiple sites, a separate site key is needed per site.