Most fingerprint databases
- Label fingerprints from monitored traffic or crowdsourced submissions.
- Cannot prove which software, version, or OS produced an entry.
- Go stale as clients update.
Database
One database for everything that speaks TLS.
Browsers, libraries, bots, AI agents, malware. Every JA4 fingerprint in Byteprint was produced by running the real software, and every entry keeps its full provenance.
The difference
Captured, not collected.
Byteprint's approach
- Runs the real software in the factory, long tail included.
- Stores product, version, OS, and TLS stack with every fingerprint.
- Reruns clients on new releases and keeps the history.
Evidence quality
Every entry can be audited.
Raw captures are retained and every fingerprint is reproduced before admission. When several clients share a fingerprint, that collision is recorded as evidence instead of being forced into one label.
Private beta