National and financial identifiers
Validated card numbers, passports, UK National Insurance numbers, and India Aadhaar and PAN numbers.
Data loss prevention that reads before it sends
An autonomous DLP agent inspects every outbound email inline at the mail layer. It reads the message, deep-scans attachments including text inside images, and makes a context-aware decision to allow, hold or stop.
See it work
Every outbound message, followed to a decision
The exposure flow shows how much mail leaves your tenant, how much of it carries sensitive data, and what Himaya held or stopped.
Exposure flowMessages observed1,284outbound mail scanned
External recipients312left the tenant
Policy matched18sensitive content found
Held or stopped7exfiltration prevented
Data Loss PreventionReceivedMessageStateDecision
9:40 AMfinance@northwind.ioTo: payments@northwind.ioNo rule matched. Policy default is ALLOW.Provider acceptedALLOW
9:38 AMomar@northwind.ioTo: partner@contoso-legal.comContract draft to an approved domain.Provider acceptedALLOW
9:31 AMlena@northwind.ioTo: l.hartmann@gmail.comSTOP selected by financial.external (IBAN or card number sent externally).StoppedSTOP
9:22 AMsam@northwind.ioTo: hr@northwind.ioNo rule matched. Policy default is ALLOW.Provider acceptedALLOW
9:14 AMmaya@northwind.ioTo: vendor@acme-supplies.netHOLD selected by source.code.external. Awaiting review.HeldHOLD
How it works
From send to decision
- Step 1
Capture
Outbound mail passes through the Himaya gateway and is safely spooled before anything else happens.
- Step 2
Extract
Text is pulled from the body and every attachment. Archives are unpacked and images are read with OCR.
- Step 3
Classify
Detectors find personal data, financial and national identifiers, credentials and your own confidential terms. A language model catches what patterns miss, such as source code and health data.
- Step 4
Decide
Your policy decides, with no model in the loop. The message is allowed, held for review or stopped, and the rule that decided is named.
Capabilities
What Data Loss Prevention covers
Credential detection
Secrets and credentials in bodies and attachments are found and reported in redacted form.
Your confidential terms
Add project names, deal codes and other internal terms. The list is versioned, so every decision records which terms applied.
Source code and health data
A language model flags code and identifiable health information, and ignores passing mentions.
ITAR and export-controlled data
Recognizes ITAR markings, USML categories and export-control notices in messages and attachments, including scanned documents.
Financial data
Bank account and IBAN numbers, SWIFT codes, wire instructions, and financial statements leaving by email.
Trainable classifiers
Upload examples of your own sensitive documents, such as contracts or designs, and Himaya learns to find similar content.
Review queue
Held messages wait for a reviewer, who sees the matched rule and findings and can release or block.
Monitor mode
Record what Himaya would have done while mail keeps flowing, then switch to enforcement when you are ready.
How we build it
Designed to be trusted
- Nothing uncertain goes out. Content that cannot be fully inspected is held, never silently allowed.
- Extracted text is not stored. Only digests and lengths are kept for each part of a message.
- Mail keeps flowing. Classification runs off the SMTP path, so accepting mail never waits on it.
Frequently asked questions
Will DLP slow down our email?
Messages are accepted immediately and inspected off the SMTP path. Allowed mail is relayed once a decision is made. Held mail waits for a reviewer.
Can we try it without blocking anything?
Yes. In monitor mode Himaya records what it would have done while every message is still delivered.
What does a reviewer see?
The decision, the rule that matched and the findings that triggered it, with release and block actions.
What if an attachment cannot be read?
It is held for review. Himaya never treats content it could not inspect as clean.
The platform
Works with the rest of Himaya
Each module deploys on its own, and they share one context layer when you run them together.
- Hive MindHive Mind connects separate alerts into one attack chain
- Email SecurityAgentic email security that investigates like a team of senior analysts
- DSPMSee every data store, and what is exposed in it
- Falcon Agent & MCPYour agentic security engineer, in Himaya and in your AI tools
- Data SovereigntyKnow where your data lives, and keep it there
- ComplianceCompliance evidence that collects itself
- Inbox PostureKeep every mailbox hardened
See Himaya Data Loss Prevention in your workspace
Book a walkthrough to learn more.