Platform overview
Vaidio AI Vision 10: the current platform release
Version 10 of the Vaidio platform: analyzes recorded video too, a modular scaling architecture, new-generation GPUs, 30+ object classes.
The real-world problem
Businesses running an older Vaidio release, or comparing AI camera platforms, need to know what version 10 changes and which infrastructure it suits. This page covers the highlights and the upgrade path.
Vaidio AI Vision 10
Version 10.0.0 · released 06/2026
How it works
01
A newer platform, more flexible deployment
Runs on Ubuntu 24.04 with new-generation NVIDIA GPUs, plus ARM support for power-efficient clusters. The software splits into separate modules, so capacity grows camera group by camera group.
02
Analyzes recorded video too
Beyond live streams, schedule the AI to scan video already stored in the VMS: investigate an incident that already happened, re-analyze an old recording batch, or process cameras without pushing a live stream.
03
Unified models and licenses
The main model recognizes more than 30 object classes: person, head, motorcycle, car, truck, bus, bicycle, animal, smoke, fire, backpack, suitcase. Custom-trained model sets exist for specific industries, and licenses are managed in one place.
Real-world applications
Upgrading on existing infrastructure
Upgrades from earlier Vaidio versions: cameras and licenses stay, enabled models carry over.
Investigating stored video
After an incident, schedule a scan of VMS recordings for the window under investigation instead of hand-reviewing tapes.
Scaling multi-site systems
The modular architecture fits large camera counts: add server groups as channels grow, licenses managed centrally.
Power-efficient clusters
ARM support (NVIDIA Grace CPU) for low-power deployments, alongside new-generation NVIDIA GPUs for compute-heavy ones.
What you get
Existing cameras become searchable data
Cameras connect over ONVIF or RTSP and recognition results store as metadata, filterable by camera, time, object type and color.
Faster investigations after an incident
Search by conditions instead of hand-reviewing tapes; scheduled scans cover recorded VMS video, with results in the same data store as live cameras.
Scaling camera group by camera group
The modular architecture adds servers as channels grow without replacing the whole system; licenses managed in one place and compared across sites.
Upgrades keep the current infrastructure
Upgrades from version 5.0.0 and later carry cameras, licenses and enabled models through the process.
How deployment works
Step 1
Survey cameras and network
Review existing cameras, VMS recorders, bandwidth between sites and the analytics needed per area to pick the configuration.
Step 2
Pick the deployment shape
On-premises Core server, multi-node clusters for large systems, edge devices for sites with few cameras, ARM for power-efficient clusters.
Step 3
Install and connect cameras and VMS
Install Vaidio AI Vision 10 on Ubuntu 24.04 with NVIDIA GPUs; connect cameras over ONVIF or RTSP and integrate the 30+ VMS/NVR platforms already in use.
Step 4
Enable analytics and operate
Load licenses online per site, switch engines on per camera and train operators; multi-branch systems continue under Command Center.
Infrastructure requirements: Existing IP cameras connect over ONVIF or RTSP with 30+ VMS/NVR integrations; analysis runs on on-premises servers or edge devices, and current cameras and recorders stay.
Frequently asked questions
What exactly is Vaidio AI Vision 10?
The current major release of the platform (10.0.0), bringing recorded-video analysis, a modular architecture and new hardware support.
Can we upgrade from an older version?
Yes, upgrades from earlier Vaidio releases are supported, keeping cameras, licenses and enabled models.
Do older cameras and recorders work with version 10?
Yes. The platform connects over ONVIF/RTSP as before; we verify resolution and angles for the chosen features.
How is crowd counting different in version 10?
The head-counting model is improved for dense crowds; see the Crowd detection feature page for details.
What does the version 10 AI model recognize?
More than 30 object classes including person, head, motorcycle, car, truck, bus, bicycle, animal, smoke, fire, backpack and suitcase, with industry model sets available.
Related features
Put AI Vision to work in your operations
Turbo Solutions surveys your current system, demos against your real operating scenarios and quotes by scope of work. No disruption to daily operations.
Book a demo