"Long lines lead to balking, waiting-line defections, and poor customer service."
Once arrivals pass ~75% of capacity, queues grow fast and customers drive off. You never see the cars you lost.
— Car wash operators' forum
Built for Malaysian hand car washes · Works with your CCTV
Your cameras already watch every car. Make them count.
WashLens reads each car's plate on the CCTV you already own, times every wash and vacuum to the minute, and shows customers how busy you are — before they drive past your queue to a competitor.
No new hardware
Runs on-site, footage stays yours
Live in an afternoon
~4 minestimated wait for a new arrival
3 / 4 bayswash bays in use right now
WVC 1234
Premium + Wax · bay 3
WASHING
12m
BKV 9871
Standard Wash
VACUUM
4m
PKD 55
Express Wash
QUEUE
1m
Avg wash today 26 min
3 cameras · updated live
A product by Terraform AI — practical AI that does the real work, human always in the loop
The problem
You know today felt busy. You don't know what it cost you.
Hand car washes run on feel — no ticket system, no timing, no data. The industry numbers say feel isn't enough.
Owners "track revenue but not labour productivity."
Wash times drift from 25 to 40 minutes on a short-staffed day and nobody notices until customers complain.
— Industry consultant
"How long will it take?" — every customer, every day.
Gyms answer with a live crowd meter. Google answers with Popular Times. Your car wash answers with a shrug.
— The gap WashLens fills
How it works
Three cameras. One timeline per car.
WashLens runs on a small box at your shop and watches the cameras you already have. No tickets, no tags, no staff input.
Reads the plate
The entrance camera reads each plate as the car pulls in — that's the start of its session, and repeat customers are recognised on arrival.
Times every stage
Bay cameras clock the car into wash, then vacuum. Queue, wash and vacuum time — measured to the minute, for every car, automatically.
Shows a live meter
You get cars-on-site, bay use and revenue. Customers get a live Quiet / Steady / Busy / Packed meter with an estimated wait.
Answers questions
Ask "which hour was busiest?" or "any car that took too long?" and the AI analyst answers from today's real data — in plain language.
One session per car — start to finish
1
Entrance
Plate read on arrival — WVC 1234 checks in. The session clock starts.
2
Wash & vacuum bays
Bay cameras time each stage — queue, wash, vacuum — with no staff input.
3
Dashboard & meter
Owner sees the numbers; customers see how long to wait. Everything updates live.
Real engine, real plates
Not a mock-up — the actual pipeline output
The WashLens engine reading real Malaysian plates from parking-lot photos and moving KL traffic. Detection and OCR run in about 10 milliseconds per frame on a Mac mini-class box.
Read WB 7255 R off a moving car in KL traffic99%
BBB 8888 — dim basement, phone quality98%
SU 3388 H — stylised private plate99%
HWE 5839 — clean daylight entrance100%
WC 6414 T — a van among many vehicles100%
TCA 3128 — picks the right plate, ignores the other94%
Every read above is live engine output on real Malaysian plates — across moving traffic, low light, and stylised plates. In the two demo clips alone, WashLens read 18 distinct plates end-to-end.
Why a lightweight camera-only tool
Figures from industry sourcesRM25–45
revenue lost per walk-away wash
~10ms
per frame — detection + OCR, on-site
$30k+
cost of incumbent tunnel LPR systems
RM0
new hardware if you have CCTV + internet
Pricing
Priced for a hand wash, not a US tunnel chain
Incumbent LPR systems are US$30k+ POS bundles. WashLens is camera-only intelligence at shop-lot prices.
Meter
Customer-facing crowd meter on one camera.
RM199/mo
- Live crowd meter page (QR + link)
- Cars-on-site counting
- Popular-times chart
Most popular
Operations
Full dashboard, up to 3 cameras.
RM399/mo
- Everything in Meter
- Plate-level session timing
- Owner dashboard + daily report
- Ask WashLens AI analyst
- Repeat-customer detection
Chain
Multi-branch, with API access.
Custom
- Everything in Operations
- Multi-branch comparison
- CRM / POS integration API
- Staff productivity module (early access)
FAQ
Questions owners ask us
Usually no. WashLens works with standard CCTV (RTSP) feeds. The entrance camera should see plates at roughly 45° or straighter and be readable at night — we check your existing footage for free before you commit.
WashLens stores plate numbers and timestamps — data you can already see with your own eyes on your own premises. Footage isn't uploaded; recognition runs on-site. Retention defaults to 90 days and plates can be hashed for analytics-only mode, with a signage template provided.
On clear entrance footage, detection and reading typically clears 95%, with Malaysian-format validation and multi-frame voting cleaning up the rest. Cars we can't read still get counted for the crowd meter — you lose the plate label, not the car.
No — WashLens deliberately doesn't do bookings, invoices or memberships. It produces the operational data your CRM doesn't have (per-car timing, live load) and can push events into your CRM via API.
A password-protected dashboard fed by a realistic full-day simulation of a 4-bay KL hand wash, in the same event format a live site produces — plus the real DeepSeek analyst answering questions about the data. A pilot install swaps the simulation for your cameras.
Half a day on-site: point the cameras if needed, install the WashLens box, draw your zones (entrance, wash bays, vacuum) on the camera view, done. You'll see your first timed cars the same afternoon.
See a full day at the wash — live, right now
The demo shows the owner dashboard and the customer crowd meter, with the AI analyst answering questions about the day.