The problem
Booking a cleaner meant chasing quotes, haggling over the price, letting in someone you couldn't check, and paying before the work was done.
My co-founder and I built Klina under our company, Gethsemane Tech, to remove all four. Verified professionals, one upfront price, and payment held until the job is done. Book in the iOS app or on the web.
- No haggling
- No pay until it's done
What I asked first
Before writing any code, we answered the questions that stop people booking a cleaner online.
- Who's at the door? Every professional is identity-verified before they can accept a single job. No anonymous strangers.
- What if the job isn't done properly? Payment sits in escrow and is released only when the customer confirms they're happy.
- How do you price without haggling? One all-in price before booking, set by our pricing system and held while they decide.
- Is the price fair to both sides? It moves with demand, within a floor that protects the pro and a cap that protects the customer.
- How do people share access notes safely? Messaging in the app. Gate codes and parking notes, without handing over a personal number.
- What if something feels wrong on a job? Every active booking carries a safety alert, one tap away.
How it runs
- Person
- AI
- Stored
- Check
- Result
Book
01, Human:
Pick the clean
Customer
Service, rooms, time and address.
02, AI:
Price the job
Pricing model
One all-in price, held for a few minutes.
03, Data:
Pay into escrow
Escrow
The pro isn't paid yet.
Match
04, Logic:
Match a verified pro
Matching
Someone nearby, usually in minutes.
Job
05, Human:
Do the job
Professional
Tracked live. Messages stay in the app.
06, Human:
Confirm it's done
Customer
The customer says they're happy.
Payment
07, Output:
Pay the pro
Escrow
Released automatically.
The pricing I built
Klina's price moves with the market, like ride-hailing, without the surprise. Rules set a base from the job. A model trained on our bookings adjusts it for everything else going on at that moment.
How a Klina price is built
- Base The job itself: service, bedrooms, bathrooms, hours and extras.
- Demand How many bookings are coming in for that area and time.
- Supply How many verified pros are free nearby.
- Timing Time of day, day of the week, and how soon it's needed.
- Place Where the home is, and how far the nearest pros are.
- Weather Conditions on the day.
- Price One all-in number, held for a few minutes while the customer decides.
Two signals tell us it's working: how quickly a verified pro accepts the job, and whether pros' earnings stay steady and above the floor.
Enforced, not requested
- A floor for the pro A price never pays a professional below a minimum.
- A cap for the customer Demand can only raise a price so far.
- The quote holds The price is locked for a few minutes while the customer decides.
- One all-in price No multipliers, call-out fees or surprises at the door.
- Verified before the first job No professional accepts a job until their identity is checked.
- Paid only when it's done Escrow releases only after the customer confirms.
- Numbers stay private Customers and pros message in the app.
- A safety alert on every job Help is one tap away on every active booking.
Inside the app
In numbers
2
countries live: the UK and Nigeria
6
kinds of clean bookable today, from weekly visits to after-builders
£0
charged until a verified pro accepts the job
2
ways to book: the iOS app and the web
6
more home services in development, from laundry to window cleaning
What it refuses to do
- Pay a pro before the job is done The money stays in escrow until the customer confirms.
- Send an unverified pro Identity is checked before a first job.
- Change a price while it's held The quote holds while the customer decides.
- Share a phone number Everything goes through in-app messaging.
What I took from it
Trust isn't a badge on a homepage. It's how the product works: who can take a job, what the price is, and when the money moves.