JA4+ network fingerprint family
A newer family of fingerprints for TLS, HTTP, TCP and more, built to survive randomization.
In one line
Can recognize your browser or app from several quiet network clues, even when shuffled
A newer family of fingerprints for TLS, HTTP, TCP and more, built to survive randomization.
JA4 (FoxIO, 2023) sorts ciphers and extensions so Chrome's shuffle gives one value. JA4S (server), JA4H (HTTP), JA4L (latency), JA4X (cert), JA4SSH and JA4T (TCP) extend it. JA4 is BSD licensed; the rest use FoxIO License 1.1.
What it can and can't do
What the user does: Nothing. It happens when your device connects and sends requests
It can
- Build a sorted code from how your browser starts a connection, so shuffling does not hide it
- Add clues from the web request, connection timing and certificates
- Use the codes to block bots and hunt for attacks
- Keep tracking the same client across different internet addresses
It cannot
- See who you are by name or account
- See page content or what you type
- Be switched off from your browser settings
- Tell apart two people with the same software and similar setup
Browser support
| Browser | Status | First version | Released | Note |
|---|---|---|---|---|
| Chrome | Supported | 1 | 2008-12-11 | None |
| Edge | Supported | 79 | 2020-01-15 | None |
| Firefox | Supported | 1 | 2004-11-09 | None |
| Safari | Supported | 1 | 2003-06-23 | None |
| Safari on iPhone | Supported | 1 | 2007-06-29 | None |
| Chrome on Android | Supported | 1 | Not listed | None |
| Samsung Internet | Supported | 5.0 | 2016-12-15 | None |
How a site asks for it
| API | FoxIO-LLC/ja4 parsers (Wireshark, Zeek, Suricata, Cloudflare) |
| How access is granted | none |
| What the user sees | Not listed |
| Needs a secure context (HTTPS) | No |
| Needs a user gesture | No |
| How long it lasts | Not listed |
| How to take it back | Not listed |
| Inside an iframe | Not listed |
Typical flow
- Parse handshake
- Sort fields
- Hash into a readable a_b_c string
Data you get
What comes back: Fixed length human readable string
Shape: string
| Field | Type | Unit | Range | Precision |
|---|---|---|---|---|
ja4 | string | None | t13d1516h2_xxxx_yyyy | exact |
Technical limits
- Spoofable by uTLS and curl-impersonate
- Software versions share hashes
The story
What it is
A newer set of network fingerprints that still names your software, even when browsers shuffle their handshake.
For you
You never see this. Many layers of your connection, from secure handshake to web request, leave small patterns. Together they point to what kind of device and software you use.
For makers
Makers use these codes because they stay useful after browsers randomize order. Defenders share lists of good and bad codes to catch bots and malware. The same data can quietly help sort visitors without cookies.
Where you see it
| Type | Example | Apps |
|---|---|---|
| security | A bot defence service sees a login attack that claims to be Safari on iPhone. The network codes match a Linux scripting tool, so it blocks the traffic. | Cloudflare |
| ads | An ad exchange flags fake traffic farms because thousands of visits share one rare network code but claim to be many different phones. | None named |
| games | A cloud gaming sign in page rejects automated account sellers whose connection pattern matches a known tool and not a real browser. | None named |
| shopping | A ticket seller keeps a list of codes seen in past scalper attacks and rejects new requests that match them. | None named |
| enterprise | A company's security team uses the codes to spot a stolen login coming from a very different kind of device than the owner's laptop. | None named |
| analytics | A threat research team groups web attacks by shared code to find which ones come from the same toolkit. | None named |
| banking | A bank adds the code to its risk score at login. A match with the customer's usual code lowers friction for them. | None named |
Risks
- Hard to see or avoid
- Unfair blocks for odd setups
Good uses and abuse
Legitimate uses
- Bot management, threat hunting (Cloudflare, AWS WAF, Fastly)
Known abuse
- Client tracking
Scores
| Score | Value | Meaning |
|---|---|---|
| Sensitivity | 2 of 5 | How revealing the data is |
| Utility | 5 of 5 | How useful it is to honest apps |
| Friction | 0 of 5 | How much the user has to do |
| Fingerprint bits | 3 | Rough identifying entropy, when known |
History
- 2023-09, tool: JA4 and JA4+ announced by FoxIO
- 2024-01, tool: Cloudflare exposes JA4 in Bot Management
Related capabilities
Related
- JA3 TLS client fingerprint. Can guess which browser or app you use from how your connection starts
- TCP/IP passive OS fingerprint (p0f, JA4T). Can guess your operating system from how your device opens a connection
Sources
How well checked: Based on official documentation.