ANSVIS turns any camera into a real-time AI sensor — detect people, vehicles, faces, plates, fire and intrusions, apply rules that cut false alarms, and trigger any system. Every frame is processed on-site.
How it works
Every ANSVIS task runs the same pipeline. Tune any stage, or leave the defaults — no code either way.
Ingest RTSP, RTMP, USB, onboard or simulated camera streams.
Crop, enhance and white-balance each frame before analysis.
Built-in or custom AI models find the objects of interest.
Zones, size, time, count and direction cut out the false alarms.
Each hit is logged with a snapshot, result and timestamp.
Fire HTTP, MQTT, VMS, email, SMS or a control system.
Prebuilt AI tasks
Pick a task, point it at a camera, press start. No training, no code — the models ship built in.
Count and track people, flag loitering or presence in a zone.
Detect people or vehicles crossing a boundary or entering a restricted area.
Detect and classify cars, trucks and motorbikes for traffic and access.
Read plates with ALPR, plus vehicle colour — entry, exit and watchlists.
Match faces to a registered gallery, with age, gender and emotion attributes.
Spot fire and smoke early, anywhere in view, and alert instantly.
Read text, barcodes and QR codes off objects as an extra pipeline step.
Detect animals for property, agriculture and wildlife monitoring.
Any camera in, any system out
ANSVIS is a desktop service that runs 24/7 on your own PC or edge box, pulling from any stream and pushing results wherever you need them.
Smart rules & detection zones
Raw detections become useful events only after your rules. Draw zones, set conditions, and ANSVIS filters out the noise before it ever reaches you.
Build your own AI tasks
When a built-in task isn't enough, compose your own in the same no-code builder. Bring a model you trained in ANSTS and wire up exactly the behaviour you need.
See it running
Real screens from the ANSVIS client — manage tasks, pick models, review evidence and read the trends.
Add, start, stop and monitor tasks — each with its camera, model, status and process time at a glance.
Use built-in models or import your own from ANSTS — ONNX, OpenVINO or TensorRT, auto-optimised for Intel or NVIDIA.
Every event keeps its image, detection result, timestamp and process time — searchable and exportable.
Summarise events by object and time in pie and time-series charts, compare tasks, and export a report.
Integrate & trigger
Every event can fire a trigger. ANSVIS speaks the protocols your stack already uses — no middleware to build.
POST event data to any REST endpoint or in-house application.
Publish events over MQTT, RabbitMQ and Pulsar in real time.
Forward events to NX Witness, Luxriot and other video-management systems.
Drive test rigs and hardware I/O through native ANS LabVIEW Integration.
Notify the right people the moment an event fires.
Alert phones instantly through the ANSVIS mobile app.
Play a sound or siren on the spot the instant a rule is met.
Trigger gates, relays and downstream automation on an event.
Why ANSVIS
Frames are analysed on-site with Intel or NVIDIA GPUs — insights without the round-trip.
No cloud, no video leaving the building. Sensitive footage stays fully under your control.
RTSP, USB and onboard cameras in; dashboards, databases, VMS and control systems out.
Rules by zone, size, time, count and direction turn raw detections into events worth acting on.
LabVIEW, VMS (NX Witness, Luxriot), MQTT, RabbitMQ, Pulsar, HTTP, email and SMS.
Run many tasks per server and manage multiple servers from one client, licensed to your needs.
Built for real deployments
Beyond detection, ANSVIS is built to sit inside a live operation — many cameras, many people, around the clock.
One server, three clients — full control on desktop, plus web and mobile views.
Add and manage several ANSVIS servers from a single client.
Keep every event with its image, then filter, review and export it.
Run tasks 24/7, on a timetable, or only when an input condition is met.
Push event images to the cloud when you want them off-site — your choice.
Control who can view feeds and who can configure the system.
Where teams use it
One wall, many jobs — crowd, ALPR, intrusion, loitering, helmet checks, wrong-way, red-light and more, live at once.
Read plates and vehicle colour, count flow, and manage entry and exit at gates and car parks.
Catch fire and smoke early and alert the right people the moment it appears on any feed.
Draw a boundary and get alerted the instant a person or vehicle crosses into a restricted zone.
Recognise registered staff and visitors, and flag unknown faces at doors and lobbies.
Count footfall, watch queues and understand busy periods across stores and venues.
Watch lines and equipment, and trigger LabVIEW or control systems when a condition is met.
Tell us the cameras and the outcome you're after — we'll show ANSVIS detecting, filtering and triggering live, on your own site.