Cloud VMS: surveillance software with no server on site

A cloud VMS takes video from your cameras over the internet and handles the rest — recording, artificial intelligence, alerts and access from anywhere — with no server to buy or maintain. Xeqmate is a cloud video surveillance platform that also runs on-premise when the operation demands it.

No server to maintain RTSP · RTMP · P2P Server-side AI Privacy by design
Concept

What is a cloud VMS?

A cloud VMS is a Video Management System that runs on servers on the internet instead of on a computer or recorder installed on site. The cameras send video over the internet, and the platform records it, runs AI on it and delivers live view and recordings in a browser and in mobile apps. There is no server, no dedicated DVR and no local software to install and maintain.

Cloud is the most common flavour of VMS SaaS: the software is contracted as a service, on a subscription, and the infrastructure is the vendor’s responsibility. If the acronym VMS is new to you, the cloud video surveillance page explains the same idea in plain language. Here the focus is how a cloud VMS works end to end — from the camera to the operator screen.

Architecture

How a cloud VMS works, end to end

The path of the video, from the camera lens to the alert on screen — with no hardware swap.

1 · Streaming

How the video reaches the cloud

IP cameras, DVRs and NVRs of any brand join the platform through three routes — with no hardware swap.

RTSP — the platform pulls the video

The cloud server connects to the camera address and requests the stream. It requires the camera to be reachable over the internet (static IP, or DDNS with a port open). It is the standard protocol for professional IP cameras, DVRs and NVRs — and the two-way route, carrying PTZ commands. ONVIF cameras come in this way: ONVIF discovers the URL, the video travels over RTSP.

RTMP — the camera pushes the video

The direction flips: the platform generates a dedicated URL, you set it on the camera or DVR, and the device starts streaming to the cloud. Because the connection is outbound, it works behind CGNAT, dynamic IPs and a closed router — with no port forwarding at all.

P2P — through the manufacturer tunnel

For supported manufacturers, the camera is registered by serial number and a P2P-enabled server pulls the video through the manufacturer tunnel — also without a public IP. For everything else, enabling the camera RTSP does the job.

One point that deserves planning: cameras consume upload bandwidth all day long — the video always leaves the site, in RTSP as much as in RTMP. Before connecting many cameras to one link, add up their bitrates and check they fit comfortably in the contracted upload. That sizing is part of what we review during the demo.

2 · Processing

Server-side AI — any camera becomes a smart camera

In a cloud VMS the artificial intelligence runs on the server, not on the camera. Any camera with an RTSP or RTMP stream gets AI, with no special hardware.

  • Video analytics: motion, person, vehicle, object, fire and PPE detection — up to 2 algorithms per camera, with detection zones drawn over the image and configurable sensitivity and operating hours.
  • License plate recognition (LPR): legacy and Mercosur plates, alerts, searchable reads and reports for co-occurrence, convoys and traffic flow.
  • Facial recognition: real-time face monitoring, face alerts and searchable face reads.

The analytics detail, with calibration and trigger conditions, is on AI video surveillance.

3 · Storage

Cloud recording with plan-based retention

Every camera has a recording plan: the combination of resolution and retention days (7, 15 or 30, for example) that defines what is kept and what it costs. Video older than the retention window is discarded automatically, and billing is pro-rated by days of use. Need to keep a clip forever? Export it — the clip stays available for download for 30 days and can be archived outside the platform.

Because the footage lives on servers away from the monitored site, a theft or damage to the equipment does not take the recordings with it — a meaningful practical difference from the DVR sitting in the lobby. The plan, retention and export details are on cloud video recording.

4 · Access and operation

Multi-platform, playback and permissions

Access from anywhere

Web (app.xeqmate.com) plus Android and iOS apps. Custom video walls, tabs, patrol cycling, PTZ with presets and tours, a camera map and a public link to embed a camera on a website.

Playback and clips

Recordings on a timeline, smart search across AI events and clip export for download — the clip stays available for 30 days to download and archive.

Users and permissions

Each user sees only the sites, cameras and modules the administrator releases, with roles, two-factor authentication (2FA), session management and a complete activity audit trail.

Security and privacy

Video in the cloud is sensitive data — and is treated as such

  • Privacy by design: lawful basis, purpose limitation, transparency and data-subject rights were part of the platform design, following GDPR, LGPD and CCPA principles.
  • Permission-based access: each user sees only what the administrator releases.
  • Audit trail: queries, alerts, administrative actions and API calls are all logged.
  • Retention under the customer control: you decide how long footage and reads stay available.
  • Two-factor authentication: 2FA available for users, with active-session management.
An honest call

When cloud wins — and when on-premise wins

Not every operation belongs in the cloud. Xeqmate offers both models on the same platform, so the answer can be an honest one.

Cloud wins when…

You do not want to buy or maintain a server; the operation is spread across several sites; scale changes often (cameras come and go); and you want speed — in the cloud model the first cameras go live during the demo itself.

On-premise wins when…

The camera count at a single site is large; internal policy requires video to stay inside the organisation; or the site upload bandwidth is too limited to send every stream to the internet. In those cases the recording server stays on site — with the same interface and the same modules.

At Xeqmate, cloud and on-premise recording servers live on the same platform, priced separately by type — you can start in the cloud and move a large site on-premise later, or run both at once. The full decision guide is in On-premise or cloud VMS: how to choose.

FAQ

Questions about cloud VMS

VMS SaaS describes the commercial model: software delivered as a service, on a subscription. Cloud VMS describes the infrastructure: processing and recording run on servers on the internet. A VMS SaaS is usually cloud-based, but — as at Xeqmate — the same service can run on-premise servers. See the VMS SaaS page.

No. IP cameras, DVRs and NVRs of any brand join the platform over RTSP, RTMP or P2P. ONVIF cameras come in over RTSP — ONVIF discovers the stream URL and the video travels over RTSP. The AI runs on the server, so no special cameras are needed.

Yes, over RTMP: the camera opens an outbound connection and pushes the video to the platform, with no public IP, DDNS or open router port. For supported manufacturers, the platform P2P route also works without a public IP.

For as long as each camera recording plan defines — 7, 15 or 30 days, for example. Video older than the retention window is discarded automatically. Exported clips stay available for download for 30 days and can be archived outside the platform.

The load is on upload and depends on the resolution, FPS and bitrate set on each camera — the stream is continuous, even with nobody watching. That is why sizing the site link is part of the conversation before connecting cameras. When the site bandwidth cannot take it, on-premise is the alternative on the same platform.

Xeqmate offers on-premise recording servers: video is recorded and processed inside the customer infrastructure, with the same interface, the same modules and the same dashboard — even alongside cloud cameras in the same operation.

The AI runs on the server: any camera with an RTSP or RTMP stream gets person, vehicle, object, fire and PPE detection analytics, plus license plate and facial recognition — with no special hardware at the edge.

Live demo

See the cloud VMS running on one of your cameras

30 minutes with our team: we connect one of your cameras live to the cloud, you watch recording and AI run, and you leave with the cost for your own scenario.

We reply within one business day · no credit card, no install, no commitment