Know what's in every file.Trust the outputs built on them.

DataXray is the unstructured intelligence layer for the enterprise. It opens every file, reads what is inside and classifies the content, then acts on it: applying labels and retention, redacting sensitive information, and feeding the results into the governance, security and AI tools you already run. It runs on-premises, in cloud or fully air-gapped, and deploys in hours.

Book a demo Race DataXray

01DiscoverEvery file and its metadataSee stage
02ReadThe content, any file typeSee stage
03ClassifyAgainst your policiesSee stage
04ActFixes what it findsSee stage

As featured in

BARC
CDO Magazine
Bank Info Security
Forrester
Breaking Defense
CIO Journal (WSJ)
BigDATAwire
Government Technology
Cybernews Live
American Banker
Infosecurity
SmartBrief on Risk and Compliance
VentureBeat

The risks

Three forces making ungoverned files an immediate liability

Catalogs govern metadata. DLP watches traffic. SIEM logs events. IAM controls access. But the files that hold the institutional knowledge your enterprise runs on sit invisible to every tool in the stack.

The AI mandate
29%of data leaders fully know where their AI-relevant data lives.
79% are confident they can use files for AI without compromising governance. The confidence is running well ahead of the evidence.
Governance by hand
74%of data leaders say governance still relies on manual effort.
Manual review doesn't survive a data subject access request (DSAR) backlog, a retention schedule, or an audit. Regimes now arrive faster than headcount.
Security exposure
36%of data leaders fully know where their sensitive data resides.
You can't protect, label, or withhold what you haven't found.

Figures from BARC, Harnessing Unstructured Data for AI Innovation, 2026, co-sponsored by Ohalo. 225 data leaders surveyed in North America and Europe.

How it works

Discover, read, classify, act

DataXray connects to file shares, SharePoint, OneDrive, cloud object storage and databases through 55+ native connectors, plus a universal connector where no native one exists. It discovers each file, reads its contents and classifies it against PII, PCI and your own sensitivity categories, then pushes that classification into the labeling, retention, redaction and protection tools you already run.

Explore the platform

Without / with

Most tools detect and alert. DataXray classifies, then acts

Detection tools report where sensitive data probably lives. DataXray opens each file and determines what is actually in it, then maps that classification directly into encryption, DLP, retention, redaction and access decisions, so a file that is flagged is a file that gets protected. Classification accuracy is 98.7%, measured on document-level PII/PCI detection in text-based English files.

Without
Sensitive data is inferred from filenames, folders and metadata. Sampling fills the gaps.
With DataXray
Every file is opened and read. Classification is based on what is actually in it.
Without
Findings arrive as alerts. Someone has to act on each one.
With DataXray
Labels, retention, redaction and attributes are applied at the file, then pushed to the tools that enforce them.
Without
Content leaves the environment for a vendor-hosted index.
With DataXray
Runs on your infrastructure, including fully air-gapped. Your data never moves.
Without
Classification that cannot explain why a file was flagged.
With DataXray
98.7% classification success rate on document-level PII and PCI in text-based English files, with the evidence attached.

See how DataXray compares

Customer story · Financial services

How a Tier-1 bank delivered file-level proof under regulatory examination

A multi-billion dollar divestiture was at risk. Three regulators wanted content-based evidence for every sensitive file, and the deadline was measured in weeks.

files classified in two months: 19M
in manual costs avoided: $22M
regulators satisfied: NYDFS, OCC and the Federal Reserve: 3

Figures from the global bank divestiture case study, 2022.

“We needed quality, flexibility, speed and proof, and we needed it in weeks, not months.”

Data Governance Director · Tier-1 Global Bank

Deployment: On-premises. Data never left the bank
Integrated with: Collibra and Microsoft Purview

Read the story

Race DataXray

10 files. You against DataXray

Read 10 files and decide which contain sensitive data. DataXray starts when you do, reading content at 100,000s of words per second, which is around 1,200 files a minute or 72,000 files an hour, across petabyte-scale estates.

The files look real, the data is fictional. Some are obvious. Some look routine until you read the third line.

Deployment

Runs where the data is

On-premises, private cloud or fully air-gapped. Files are read in place. Content never leaves your environment, and there is no vendor-hosted index.

Deploys in hours. Containers on your infrastructure, read-only connectors to the sources you already have.

Built to enrich the stack you have, not replace it. File-level attributes flow to Microsoft Purview, DLP, encryption, SIEM, and catalogs. Where no connector exists, the API does.

How DataXray deploys

Enterprise and government

The certifications your vendor security review will check

DoW ATOAuthority to Operate on three networks in the U.S. Department of War
SOC 2 Type 2Ohalo Limited, annual audit
ISO 27001Information security management
FIPSFIPS-compliant cryptography
STIGSTIG-compliant implementations

Explore security and trust

Modules

Built on every file DataXray has already read

Modules extend the DataXray platform into the workflows customers ask for most. Each one runs on files DataXray has already read and classified, so no file is read twice. Every module is configured to fit how your team works.