Classifyre · Documentation
Scattered data
in. Closed cases out.
Classifyre reads the systems you already run, raises the findings worth a human's time, and gives you a structured place to work them — with an AI autopilot doing the legwork between scans. These docs explain how the application works, section by section.
Step zero — get it running
Classifyre ships in two distributions, with the same features in both. Install one, then work through the pipeline below.
Option 01 · Desktop
Install it on your machine
One installer for macOS, Windows, or Linux. PostgreSQL and the scan workers are inside the package, so there is nothing to provision — and only the desktop build can scan a local folder.
macOS · Windows · Linux — free, no signup
Option 02 · Kubernetes
Or run it on a cluster
Helm, OCI-native · no repo add
The chart deploys the web UI, API, worker, and the ephemeral scan Jobs. PostgreSQL can be embedded for a trial or external for production.
helm install classifyre \
oci://registry-1.docker.io/classifyre/classifyre-core \
--namespace classifyre --create-namespace \
--version 0.4.132Then: one pipeline, five steps
- 01ConnectPoint Classifyre at systems you already run. Data stays in place.
- 02ScanEach run registers every document and file as an asset.
- 03DetectDetectors read every asset and raise findings worth a look.
- 04Connect the dotsInquiries keep watch; fingerprints link the same thing everywhere.
- 05ResolveCases collect evidence, weigh hypotheses, and reach a conclusion.
Start where you are
New here
Understand it
The whole platform in plain English — no setup, no jargon, just how a pile of scattered data becomes three leads worth your time.
Setting it up
Run it
Stand the platform up, connect your first system, and watch the first scan land — on a laptop or in production.
Working results
Investigate
Findings are the start, not the answer. Watch them with inquiries, connect them with fingerprints, and close them as cases.
Every section
Looking for a specific connector or detector? The source catalog and detector catalog are generated straight from the product, so they are always current.