Blog / How it works
What JA4 is and why it recognises a bot
A bot can claim to be Chrome and send each request from a new IP address. What it cannot easily change is the way its software opens an encrypted connection. That is what a JA4 fingerprint describes.
JA4 is an open method for fingerprinting how a client starts a TLS connection, the successor of the older JA3. It turns the first message of every encrypted connection into a short string that is the same for the same software.
What happens when a connection opens
- ClientHello
The client sends the TLS versions, cipher suites and extensions it supports, in an order typical of its software.
- JA4
That message is condensed into a short fingerprint. A browser, a library or a script each leave their own.
- Rule
WEDOS Protection reads the fingerprint of every connection and applies the rules that match it.
Why a fake User-Agent does not help
| Signal | How a bot changes it | What that means |
|---|---|---|
| User-Agent header | One line of code | It says whatever the client wants |
| IP address | Rotating proxies | It changes with every request |
| JA4 fingerprint | It needs a different TLS stack | It stays the same across IP addresses |
A script written with a common HTTP library can send a Chrome User-Agent, but its connection still opens the way that library opens it. The fingerprint gives it away, even when the requests come from thousands of addresses.
Where it can mislead
A fingerprint identifies software, not intent. Many legitimate tools and apps share the same libraries, so a rule on JA4 alone can catch more than you want. It works best combined with other conditions, such as the path, the country or the HTTP method, in one combo rule.
In WEDOS Protection
JA4 is built into WEDOS Protection. You can write rules by fingerprint, use JA4 in combo rules together with other conditions, and with advanced fingerprint matching also look at the JA4H fingerprint of the HTTP request itself.