Computer vision for UK shops, cafés, and pubs.
The camera above your door has been counting people for years. You just haven't seen the numbers. I build vision systems that turn your existing cameras into footfall counters, queue alerts, shelf checkers, and till aids. Built on what you already own. Paid once, yours forever. No monthly subscription, no vendor lock-in.
Or copy: [email protected]
What's actually happening in your shop right now.
You're guessing. You think Tuesday afternoons are quiet. You think the queue at lunch is fine. You think the bread shelf gets refilled every morning.
The chains know all of this in real time, from the same cameras you already have above your tills and doors. They pay a per-shop subscription for a product that gives them the numbers, on a dashboard locked to the vendor.
You can have the same data, from the same cameras, paid for once. The code lives on your hardware. The system is yours.
What I build for shops, cafés, and pubs.
Footfall & dwell counter
How many people came in, when, where they lingered, conversion to till. Replaces the per-shop people-counter subscription with something you own outright. Daily summary email lands first thing: yesterday's traffic, peak hour, slowest hour, and a comparison to the same day last week.
Quoted on scope.
Fixed price in writing before any work starts. Hardware, on-site installation, and any opted-in subscriptions quoted separately at cost.
Queue length alerts
Tells staff (or a screen in the back) when the queue hits a threshold you set. Cuts the moment when three customers walk out because nobody noticed the line. Works on tills, bar service, takeaway counters.
Quoted on scope.
Fixed price in writing before any work starts. Hardware, on-site installation, and any opted-in subscriptions quoted separately at cost.
Shelf gap & stock-out detection
Spots empty shelves from existing cameras. Alerts a phone or sends a list to whoever fills shelves at the start of the next shift. Best for high-velocity lines: bread, milk, beer, the things that walk out fast on a busy day.
Quoted on scope.
Fixed price in writing before any work starts. Hardware, on-site installation, and any opted-in subscriptions quoted separately at cost.
Age-restricted sales aid
Flags till transactions where the customer may need ID, based on age estimation from the till camera. A prompt on the screen, before the sale completes. Not a replacement for staff judgement, just a quiet backup that catches the moment a tired colleague forgets to ask.
Quoted on scope.
Fixed price in writing before any work starts. Hardware, on-site installation, and any opted-in subscriptions quoted separately at cost.
A few of the situations this system handles.
A shop owner who wants real footfall numbers, not the till count. The system uses the door camera to count people in and out, breaks the data down by hour and day, and lands a one-page summary in the owner's inbox every morning.
A pub or café with a bar queue that backs up unpredictably. The system watches the queue area, sends a quiet alert to the back of house when it exceeds the agreed length, and logs peak times for staffing reviews.
A small chain with several shops under one operator. The system runs in each shop and reports back to a single owner dashboard, showing footfall, conversion to till, and any outliers worth a closer look.
A shop selling age-restricted goods at busy times. The system flags till transactions where the customer may need ID, prompting the till operator quietly before the sale completes, as a backup to staff judgement, not a replacement.
Why this, why me.
It runs on your existing cameras.
The cameras above your doors and tills are almost always good enough. Clip review tells you for sure, free.
The code is yours.
Most footfall-counter vendors charge per shop per month, locked to their dashboard, their data, their pricing. Mine ships as code in your repo. Paid once.
One person, start to finish.
You email me, you call me, you get the build from me. No account managers, no support tickets.
GDPR-safe.
Faces blurred by default. The system counts people, not who. The data stays on your hardware unless you choose otherwise.
Fixed price, in writing.
The quote is the price. No surprise invoices.
A call, then a build.
We talk
Fifteen minutes. You tell me how many sites, what you want measured, what's installed already.
I build
One to three weeks for most retail installs.
Every quote breaks out the build fee, hardware, install labour, and any subscriptions in writing before you commit. One total, every line visible.
The things people ask before we start.
Will this work with my existing CCTV?
Yes, in the vast majority of shops, cafés, and pubs. Door cameras and till cameras from any modern install will work. The cases where it doesn't are usually a fisheye lens or a very low ceiling mount, and we sort that out at the scoping stage.
Where does the data live?
On a small computer in your shop's back room. Yours. Off-site cloud setups are available if you want them, but most independent retailers prefer everything to stay in the shop, both for cost and because it keeps the data conversation simple.
Can it identify customers?
No, and it shouldn't. The system counts people anonymously. Faces are blurred at the first processing step, before any count, alert, or log is created. Facial recognition is a different piece of work, much more heavily regulated, and not something I recommend for independent retail.
Will it work in a small shop?
Yes. One camera above the door is the minimum useful setup, and most independent shops already have this. Two cameras (door and till) opens up more useful data, but it's not required.
What if the camera angle is bad?
We pick this up at the scoping stage. If the angle is too low, too high, or the lens is too wide, I tell you before any money changes hands. Sometimes the fix is repositioning what's already there. Sometimes it's a single new camera. Either way, no surprises after the fact.
Can it integrate with my till system?
Often yes. Most modern EPOS systems have an API, and we can match footfall data to till transactions to give you a real conversion rate. Older till systems are harder. We talk through your specific setup on the call.
How quickly can it be running in my shop?
One to three weeks for a single shop. Multi-shop rollouts run faster per shop after the first one, because the configuration is reused. A small chain of five shops typically completes inside four to six weeks total.
Tell me what you need built.
Email me a sentence about what you need. I reply the same day.
Or copy: [email protected]