Platform overview

MyDLP is an endpoint-first data loss prevention platform. A small agent on each computer finds sensitive data locally, decides what to do against a signed policy, and reports a redacted decision to a central server. The server turns those decisions into incidents that an analyst works in the console.

This page is the map. Each section links to the page that holds the detail and the limits.

Five design decisions

Decision	What it means for you
The endpoint decides.	Detection and policy run on the computer, against a cached, signature-verified policy. Protection continues when the network is down, content is not sent to a server to be judged, and no round trip sits in front of a file write.
The simulator is the enforcement code.	The same detector and policy engines are compiled into the agent and the server. A policy you simulate in the console is evaluated by the code the endpoint runs, so the preview cannot drift from the result.
No claim of enforcement the endpoint cannot perform.	Every agent reports which actions each channel supports. A policy action is reduced to what the endpoint can do, and the incident records what was asked and what was applied.
Matched values never leave the endpoint.	Redaction happens in the agent before an event is built. The server receives detector, category, counts, confidence and a masked preview, and has no code path that stores a raw match.
Identity fails closed, availability fails open.	An unverified device is refused. An endpoint that cannot vouch for itself still reports, loudly.

Components

One file, end to end

The path a single file takes, from the change being seen to an incident. Every step in the left panel runs on the endpoint.

Step-by-step detail, including budgets and deterministic event ids, is in Architecture.

Detail: Architecture, and for a clustered deployment Kubernetes reference architecture.

What the platform does

Area	Summary	Read more
Detect	Built-in detectors for financial, identity, credential and organisational data, country packs, exact data match and document fingerprints	Detector catalogue
Decide	Policies with scope, conditions, exceptions, schedules, conflict rules and risk thresholds	Policy model
Protect	Channel monitors on the Windows agent: files, removable storage, cloud-sync folders, clipboard, print, browser upload, outbound connections	Capability matrix
Investigate	Incidents correlated per content, explainable risk score, evidence access with justification, legal holds	Feature catalogue
Discover	Scans of mounted shares and S3 buckets for sensitive data at rest	Feature catalogue
Administer	Roles, directory sign-in, SCIM, API tokens, audit chain, retention, update rollout	Feature catalogue
Integrate	Webhook, syslog, Splunk HEC, Slack, Teams, Jira and ServiceNow	API

What it is not

The register of limits is Known limitations. The shortest version: the production Windows agent observes and warns but does not block file operations; the Linux agent is frozen and is not documented here; no detector claims zero false positives; and nothing here claims complete DLP coverage.