A tour of EndSecure
From the first AI tool found to the report on your desk, in the order you would meet it.
This is the walkthrough we would give you on a call, written down so you can read it whenever suits you. Each stop is a part of the product that works today, shown with illustrative sample data.
Last updated: September 2026
Your security partner
Kobu will walk you through it
Six stops, one sensitive paste. If a question comes up on the way, Ask Kobu (bottom right) searches the FAQ for you.
Six stops, one sensitive paste.
Follow one event through the console. The tour moves on by itself; pick any step to jump to it.
Step 1 of 6
Find the AI already in use
The sensor and browser extension report cloud AI in the browser, desktop AI apps and local models. Each one lands in the AI inventory with its telemetry level, so you can see how much of it EndSecure can actually observe.
Read the details1Where it starts
Find the AI your people already use
Nothing else in EndSecure is useful until it knows which AI tools are running. So the first thing that happens after you install the EndSecure AI Sensor on a Windows endpoint is discovery.
The sensor looks for installed AI software, AI apps starting and stopping, and local model runtimes. The Chrome and Edge extension reports supported AI websites. Everything lands in one AI inventory, where you mark each tool as approved, unreviewed or blocked.
- Cloud AI in the browser, such as ChatGPT, Claude, Gemini and Microsoft Copilot
- Desktop AI apps and developer tools
- Local LLM runtimes such as Ollama, LM Studio, AnythingLLM, LocalAI, Jan and llama.cpp
- A telemetry level on every application: FULL, PARTIAL, METADATA ONLY, PROCESS ONLY or UNKNOWN
2Before anything is sent
Recognise sensitive data on the device
Where the application and the endpoint integration make content visible, EndSecure DLP checks it before it is submitted: a paste, text typed into a supported site when Send is pressed, or an uploaded file.
Detection runs on the endpoint. What reaches the platform is a masked finding: the kind of data, where it was found and how risky it is, never the full value.
- Secrets such as API keys and credentials
- Personal and customer identifiers
- Financial identifiers such as PAN, Aadhaar and IBAN
- A risk level for each finding, used by your policies
3Your rules
Decide with policies you can read
Every finding meets a policy. Rules read like sentences: when this AI tool, and this kind of data, then Allow, Warn, Block or Redact, or send it for Review or Route it to an approved tool.
You do not start from a blank page. Setup offers Balanced, Strict and Monitor only templates, and the What-if simulator shows what a change would have done to recent events before it goes live.
- WHEN / AND / THEN rules with six actions
- Balanced, Strict and Monitor only templates
- What-if simulator on past events
- Every policy change written to the audit log
4In the moment
Stop a risky submission before it leaves
When a rule says Block, the browser extension stops the paste, the send or the upload on a supported AI site and tells the person why, with the sensitive value masked. A Warn lets them edit the message or continue.
For desktop AI apps, Clipboard Guard checks pasted text against the same policy.
- Chrome and Edge extension for supported AI websites
- Checks pastes, typed text on Send and file uploads
- Clipboard Guard for pastes into desktop AI apps
- A plain explanation for the person, not a silent failure
5After the decision
Investigate with masked evidence
Where your policy asks for it, a decision opens an incident. It shows the endpoint, the AI tool and whether it was approved, the channel, the rule that fired and a timeline from the paste to the decision.
Evidence stays masked, so an analyst can understand what happened without anyone reading the original prompt. Acknowledge, assign and resolve it, and the audit log keeps the record.
- Incident list with severity and status
- Masked evidence and the rule that fired
- Timeline of the event, including the Wazuh alert if you use Wazuh
- Roles: Owner, Admin, Security Analyst and Viewer
6Keeping score
Report on AI use and the decisions made
The dashboard brings together AI applications, events, decisions and endpoint health, so you can answer the weekly questions without building a query.
Endpoints show each sensor’s heartbeat and health, and the audit log records policy changes, decisions and admin actions for the people who need to check them.
- Dashboard with AI applications, events and decisions over time
- Endpoint health with heartbeat and sensor status
- Audit logs for policy changes and admin actions
- Two-factor authentication for every account
Before you join the waitlist
Two things worth knowing plainly
The first is coverage. EndSecure runs on Windows endpoints, and content inspection happens only where the application allows it: supported AI sites in Chrome and Edge, and pastes into desktop AI apps. Hand-typed prompts in desktop apps are not inspected, and every AI application in the inventory shows its telemetry level so nobody mistakes discovery for full visibility.
The second is cost. Plans run from a first evaluation to an organisation-wide rollout, and pricing is revealed at launch. People on the waitlist see early-access pricing first.
That was the tour. Pick whichever next step suits you.
Try it yourself
Join the waitlist for early access. When your spot opens you get a workspace and the sensor for one Windows endpoint, with nobody calling you first.
Have us walk you through it
We will run the same walkthrough against your environment, your AI tools and the questions this page did not answer. Bring the difficult ones.
Still weighing it up? The comparison page sets EndSecure next to other approaches, and the security answers cover what your reviewers will ask.
Questions after the tour
What EndSecure does, what it doesn’t, and how it fits with your tools.
See all questions