Clickproof
A tool that tells real clicks apart from bots, duplicates and forwards in paid campaigns — so a business only pays for genuine traffic. We built it as a working prototype on the Vercel edge.
The brief
Which clicks are you actually paying for?
In paid campaigns on Instagram or through affiliate partners, a business pays per click — but some of those clicks aren't people. Bots, duplicate opens and in-app-browser forwards burn budget without a single customer.
Standard analytics don't show it. There was no tool that, for every click, says whether it's real — and how much of the traffic is even billable.
AI Automation
What we built
A link tracker that scores every click
We built a tool that scores every click on a tracked link in real time. A clear dashboard shows how many clicks are clean, how many are duplicates and how many are suspicious — per link and overall.
How it works
A device fingerprint and dozens of signals
For every click we take a device fingerprint and dozens of signals — headers, geolocation and ASN, click-ID age, fingerprint entropy and in-app-browser behaviour. They combine into a 0–100 quality score and a fraud flag.
- 01Signals like zero hardware concurrency or device memory expose automated environments.
- 02Geolocation, language and referrer are cross-checked — a mismatch lowers the score.
- 03Every click gets a clear verdict: clean, suspicious or fraud.
Billable vs. fake
Pay only for real traffic
The tool separates billable clicks — clean, scored above 80 and with no duplicate — from suspicious, fraud, duplicates and forwards. A business sees what it should really pay for, and partners can't over-bill it for bots.
Don't spend budget on clicks that were never a customer — a simple principle we turned into a working tool.
Results in numbers
0–100
quality score for every click
5
classes: clean, suspicious, fraud, duplicate, forwarded
real-time
every click scored on the Vercel edge
Have a similar project?
We'd love to get it moving — let's start with a no-obligation chat.