Keepvik
Keepvik · Managed security

Know exactly what your company PCs are doing on the network

A lightweight agent inspects connections, exposed services and security settings on each computer. Our analysts review every finding with you and agree the fix. AI assists — humans decide.

For Windows 10 & 11 · one-time check or ongoing monitoring

How it works

  1. 1
    Enrol

    Your IT admin runs the Keepvik agent on each company PC and enters your company code.

  2. 2
    Inspect

    It reads connection metadata, open ports, and security posture — never file contents or browsing history.

  3. 3
    Analyse

    Rules + threat intelligence flag real indicators. AI drafts an explanation for our analysts.

  4. 4
    Review with you

    Our team confirms what's real, tells you where the problem is, and agrees the fix with you.

  5. 5
    Remediate

    With your approval on the device, we apply safe, reversible fixes — or hand your team clear steps.

The product is the diagnostic engine and our analyst team. AI only drafts explanations from measured evidence — it never decides what's wrong, and it never talks to you directly.

What we check

Suspicious connections

Every TCP/UDP connection matched to the program that made it, and checked against botnet / malicious-IP threat feeds.

Exposed services

RDP, SMB, VNC and other sensitive services listening on the network — the usual ransomware entry points.

Security posture

Antivirus, firewall, UAC, SMBv1, RDP hardening, BitLocker, admin accounts.

Persistence & malware traces

Startup entries, scheduled tasks and unsigned programs running from unusual places.

Network integrity

Hosts-file tampering, proxy hijacking and unusual DNS.

Performance

CPU, memory and disk — because crypto-miners and problems hide there too.

Privacy & consent, by design

  • ✓ Runs only on computers your organisation enrols with its own code.
  • ✓ Reads connection metadata and system settings — never file contents, keystrokes, screens, or the pages employees browse (browser traffic is counted, not itemised).
  • ✓ No fix ever runs without someone approving it on the device. Every action is from a fixed, reversible whitelist — there is no “run any command”.
  • ✓ Full audit trail of every scan, proposal and action.