What you actually get, and why.
AllCars has six features that matter. None of them are there becuase a roadmap said so. Each one came out of a specific annoyance, a missed opportunity, or a car I shouldnt have bought. Here they are, in order of how much pain they removed for me first and then for everyone else.
All the ads, in one place. Deduplicated.
Every public car listing in Cyprus, indexed and merged so the same car never appears twice. One car. One timeline. One truth.
How it works
Every public car listing in Cyprus gets pulled into one index, refreshed continuously. When the same physical car appears in three places at three different prices, the deduper spots it using image fingerprints (perceptual hashes of public photos) and spec match (year, make, model, mileage, engine all close enough to be the same car). It merges them into one vehicle record with one unified timeline.
Hard veto rules stop false merges. Different body type, different fuel, different colour and the merge is rejected even if every other signal screams "match". Different cars stay different cars, always.
I got tired of opening four tabs every time someone said "have you seen the X for sale?". The same car would show up in three places at three different prices and I couldnt tell if I was looking at one car or three. So I made the engine merge them. One car. One timeline. One truth.
Every other Cyprus car site is a single source pretending to be a market. AllCars is the market. The dedup work means the same dealer relisting a car five times under three slightly different spellings counts as one car with one price story, not five fake listings inflating supply.
Offer intelligence, at a glance.
Every car gets a 0 to 100 score against its real peer group. Green for steal. Amber for fair. Red for above market or flagged. Skim a list of 200 cars in 30 seconds.
How it works
Two scores get computed for every listing. Value asks "is this priced well for what it is?". Quality asks "is this car actually what it claims to be?". The deal score is the two of them multiplied together, normalised to 0 to 100.
A car priced 20% under market with proper photos and a real description gets green. A car priced 30% under market with three blurry photos and salvage keywords in the text gets red, because cheap-for-a-reason isnt a deal. The colour code reads itself.
Want the friendly walkthrough? Score guide. Want the maths? Pricing engine.
Looking at 200 listings to find the good ones is a part-time job. The deal score collapses that into a glance. Green = look closer. Red = scroll past. Amber = use your judgement. Saves about three hours of clicking per session, and your brain doesnt go numb halfway through.
Other sites give you a "fair price" estimate based on a flat average. AllCars compares each car to its real peers, penalises listings that smell off, and shows the colour instantly in the browse view. You see the score before you click. Nobody else does that in Cyprus, and not many do it this honestly anywhere.
The pricing engine. v6.
Real comparables. Real maths. No vibes. The engine that turns every listing into a 0 to 100 score has been through six versions, each fixing something the last one got wrong.
How it works
For every car, the engine finds its real peer group across 13 dimensions (model, year, mileage, engine, fuel, body, gearbox, condition signals). The closest peers count most via a smooth weighting function. The engine works out the typical price for this exact car shape, adjusts for mileage and condition, and compares the asking price.
For rare cars with few peers, it pulls the estimate toward the broader segment using adaptive Bayesian shrinkage. For cars that smell like lemons, it flags them via the Akerlof Value × Quality gap. Every ad you see has been through all of this.
Read the full architecture pageFive years ago I bought a car for what I thought was a fair price. Six weeks later the same model went for €3,200 less. I spent the next two years building a peer-finder, a kernel weighter, a Bayesian shrinker and four other things, just so it never happens to me again. Now its yours.
This level of pricing maths exists in commercial dealer tools that cost €2,000 a month. AllCars puts it in front of buyers, in a public app, for free. Nobody else in Cyprus does this. Most fair-price tools dont even exist on this continent at this depth.
The engine watches for you.
Save a search, save a specific car. Telegram pings you the second a new match appears or a watched car drops in price. While you sleep. Whilst you work. Always.
How it works
Save any search you find yourself running twice. Same filters, same model, same price ceiling. The engine remembers and watches the index for you. New matches and price drops on watched cars trigger a Telegram message within minutes of the listing appearing.
You also get the deal score in the notification, so you can decide instantly whether to drop everything and call the seller, or keep watching telly. Most matches arrive between 7am and 10pm. The good ones go fast.
The car you want is online for about six hours before someone else clicks "buy". By the time you check tomorrow, its gone. So the engine watches for you and pings your Telegram the second a match shows up. I missed three good ones before I built this. Wont happen again.
Most car sites have "saved search" but the email arrives the next morning. AllCars notifies via Telegram in near-realtime, with the deal score embedded so you know whether its worth interrupting whatever youre doing. Plus its free and theres no inbox to drown in.
Every ad ever, kept.
First asking price, every drop, every relist, every disappearance. The full price story for every vehicle, even after the listing is long gone. The market has a memory now.
How it works
Every observation gets logged into an append-only history. Nothing is overwritten, nothing forgotten. When a listing appears, every refresh records its current price. When the price drops, the drop is timestamped. When the listing vanishes, that gets timestamped too.
Tap any car and you can see whether its been online before under a different price, whether the same dealer has relisted it three times this year, and how long similar cars typically stay on the market before they sell. Real context. Not just a snapshot.
Every listing has a story. Was it relisted? Did the price drop? Was it for sale a year ago at €4,000 more? The history layer keeps every observation forever, so I can see whether "rare find" is actually rare or has been sitting on the market for 90 days at three different prices waiting for someone gullible.
Most listing sites delete the past. The ad disappears, the history goes with it. AllCars treats every observation as permanent record. You can search for a 2018 A4 and see not just whats live now, but what similar cars sold for last month, last year, ever. 700,000+ price observations and counting.
Six features. One simple promise.
Spend your attention on cars, not tabs. The engine does the hopping, the sorting, the maths and the watching. You drive.