Classifyre

Classifyre finds what matters in the data scattered across the systems you already run — and turns it into investigations you can act on. Connect a source, let detectors surface the evidence, and work the results like an analyst: standing questions, duplicate detection, cases, and hypotheses — with an AI autopilot doing the legwork in between.

It’s an open-source investigation platform for your whole data estate: databases, lakehouses, collaboration tools, content platforms, and more.

What you can do with it


Connect what you already run	Point Classifyre at databases, data lakes, storage buckets, collaboration tools, and content platforms — no data migration.
Detect what matters	Ready-made detectors for secrets, PII, and security, plus custom detectors you build — from a simple rule to a full AI model.
See the same thing, everywhere	Fingerprints connect duplicate and similar records across systems into a single picture.
Investigate, don’t just list	Turn findings into cases with evidence, competing hypotheses, and a clear conclusion.
Let AI do the legwork	An autopilot opens inquiries, builds cases, and drafts hypotheses after every scan — explaining every move.
Run it your way	Start on a laptop with Docker, scale to Kubernetes, keep your data in your own infrastructure.

How it all connects

Classifyre is one connected pipeline — from the systems you connect, all the way to a resolved investigation:

Sources connect the systems you run and scan them into assets with rich metadata.
Detectors read each asset and raise findings — the signals worth acting on.
Inquiries keep watch over findings, and Fingerprints connect the ones that share identity.
Cases organise evidence and hypotheses into real investigations.
Autopilot does it for you, automatically, after every scan — with a written reason for every action.

The building blocks

Area	What it covers
Sources	Connect systems, choose what to scan and how, and understand what a scan produces.
Detectors	Pre-built packs and custom detectors, how they run, and the findings they yield.
Investigations	Inquiries, fingerprints, cases, and hypotheses — turning findings into answers.
Autopilot	The AI agents that run investigations for you, and how to steer them.
Settings	AI providers, instance settings, and connecting your own tools.
Deployment	Run Classifyre with Docker or Kubernetes, on your own infrastructure.

Get started

Try it — stand up the whole platform with the Docker All-in-One guide.
Go to production — deploy the open-source core with the Kubernetes Helm chart.
Connect your first source — see the Sources guide.
Understand the journey — follow a scan from start to finish in Flow.

Explore the documentation

Flow: The end-to-end lifecycle from a source scan to findings and investigations.
Sources: How sources work, what to configure, and what a scan produces.
Detectors: Pre-built and custom detection, and the findings it yields.
Investigations: Inquiries, fingerprints, cases, hypotheses, and the AI autopilot.
Deployment: Docker and Kubernetes installation, plus the platform architecture.
Settings: AI providers, instance settings, and the MCP server integration.
Data Export and Notifications: Getting data out, and staying informed.