Turn the CCTV you have
into an AI control room
No camera replacement. AI analyses your existing CCTV streams around the clock and alerts you the moment a hazard appears.
Start on the Free plan — up to 3 cameras at no cost
Automatic AI detection of fire, intrusion, falls and more
Real-time AI analysis, even on an ordinary office PC
You have cameras. Nobody is watching them.
Accidents happen in the moments no one is looking, and recorded footage only helps after the fact.
MotionCat takes your existing CCTV streams as they are, analyses them continuously, and alerts you the instant a hazard appears.
No new hardware
The office PC already handling your CCTV is enough to start. A smartphone can even serve as the camera.
AI watches, not people
Fire, intrusion, falls and other hazards are detected automatically, 24/7. Your team steps in only when it truly matters.
Alerts are automatic too
SMS goes out automatically based on severity, so the right person can respond at night and on weekends.
Scene analysis — ask the camera in plain words, get an answer
Describe the situation you care about in a sentence. The AI compares it against the live frame and tells you when it matches.
Runs from day one, no training
Built on a large vision-language model, so there is no training, data collection or setup before analysis begins.
Your own wording, many at once
Register as many situations as you like and apply a different set to each camera.
A similarity score, not a guess
See how closely the current frame matches each situation you registered, as a score.
Unlimited on every plan
Unlimited on all plans including Free — apply it to every camera from the first day.
How to write a keyword
A short sentence that paints the scene works better than a single word.
The more the sentence names a subject and what it is doing, the more precisely the model separates one scene from another.
Start with scene analysis for broad coverage, then extend the cameras that need close watch to severity analysis (next).
Severity analysis — a purpose-trained model where it matters most
We train a dedicated model on the situations your site defines, and grade risk on three levels in real time.
Three severity levels
Ask us and MotionCat builds the custom model for you.
Five business days from request to a model trained for your site.
A model we train ourselves
Case-specific rather than general-purpose, so accuracy is higher — and we keep measuring inference speed and accuracy to improve it.
Light, fast inference — 50-100ms
Real-time analysis at office-PC spec with no GPU server. Processed at the edge, with no investment in expensive hardware.
Custom cases
Add fire, intrusion, falls, equipment faults or whatever your site defines as a dedicated case (one to unlimited, by plan).
An independent score per case
See which hazard is at which level. Built-in hazard cases start analysing the moment a camera connects.
AI reports — what happened, written out by AI
When an event fires, cloud AI looks at the footage and describes the situation in natural language.
“A worker has fallen and is lying on the floor”
The summary is recorded together with the severity assessment.
Triggered automatically by severity
Runs automatically on alert (3) and above, caution (2) and above, or every event depending on your plan — and can also be requested manually.
Multilingual
Describes the situation in Korean, English or Japanese.
No load on site hardware
Heavy computation runs in the cloud. Your on-site PC spec stays exactly as it is.
Minimal data leaves the site
Only the image for that event is sent, and only when captioning is used. Turn media sharing off and no other video or image leaves your network.
Nothing new to buy
On-site installation means putting the analysis program on a PC — that is all. From registering a camera to analysis running: the same day.
On-site CCTV · smartphone
Your existing equipment
On-site PC (analysis program)
Scene and severity analysis
Cloud control console
Monitoring · alerts · statistics · captioning
Raw video never leaves the site — analysis runs at the edge, monitoring in the cloud. Your existing NVR recording keeps running untouched.
Live the same day
From camera registration to analysis
An office PC is enough
No GPU server, no dedicated appliance
Raspberry Pi 5
Runs on an 8GB single-board device
Flexible placement
Spread across several PCs as you scale
Pricing that scales with your cameras
MotionCat is licensed per camera, per month. Tell us how many cameras your site runs and we will send a quote in your currency, with terms for your region.
Start on the Free plan
Up to 3 cameras with unlimited scene analysis, at no cost and with no sales conversation required.
Per-camera monthly licence
Paid plans are priced per camera per month, and scale by how much severity analysis and AI captioning you use.
Enterprise and on-premise
Air-gapped deployments, custom models and captioning limits are set to the size of your site under a separate agreement.
Distributor and system-integrator enquiries are welcome. We hold an MOU with Japan’s OPEN Group and are expanding beyond Korea.
MotionCat — OOLU Soft Co., Ltd.
Start today
Connecting three cameras you already own is enough to begin.
See scene analysis and severity analysis for yourself on the Free plan.