Unattended access
Reach any enrolled device whether or not somebody is logged in. Servers at 3 a.m., a laptop in a hotel room, a kiosk in a warehouse. Wake-on-LAN included.
Remote Monitoring & Management
The fix runs itself. You watch the log.
UniDesk RMM puts a lightweight agent on every laptop, desktop, and server — with remote desktop, patching, and monitoring built in. Then Lyvn reads the ticket, picks the matching script, and runs it on the endpoint before anyone opens a session. When a human is needed, remote control is one click away in the same window.
14-day free trial · Windows · macOS · Linux · Unattended access included
app.unidesk.ai/devices/MBP-4471
MBP-4471
Run scriptRemote control
ScreenConsoleFilesProcessesEvent logREC 00:0042 ms
root@MBP-4471running…
Live vitals
every 30 sCPU84%
Memory77%
Disk used96%
Automation run
TKT-8812 · “My Mac says the startup disk is full”
Matched to runbook
macos-disk-reclaim · conf 0.96Local snapshot created
purge-caches.sh started · signed · rootReclaimed 33.2 GB
Health check passed
TKT-8812 resolved · replied to Olivia
Signed script · scoped credentials · session recorded
One agent forWindows 10 / 11Windows ServermacOS 12+Ubuntu & DebianRHEL & RockyVMs
The loop
Most RMMs tell you something broke. This one has already fixed it.
A monitoring console is only useful if somebody is watching it. UniDesk turns every alert and every ticket into the same thing — a record with a device attached and a runbook waiting.
01
Monitor
A lightweight agent reports CPU, memory, disk, battery, patch state, drive health, and 200+ other signals every 30 seconds. Offline devices report the moment they wake.
02
Detect
Alert policies fire on thresholds, event-log patterns, failed services, or missing patches — and each alert opens a real ticket, not a line in a console nobody reads.
03
Remediate
Lyvn matches the ticket to a runbook in your script library and runs it on the endpoint. Signed script, scoped credentials, snapshot first, full transcript.
04
Verify & close
The agent re-runs the health check, confirms the symptom is gone, replies to the employee, and closes the ticket with the run log attached. Escalates if the fix didn't hold.
Remote desktop
Take the screen in under three seconds.
Full remote control is part of the platform, not a second vendor you renew every January. Attended when the employee is there, unattended when they're not.
Attended sessions
Send a join link in Slack, Teams, or email. The employee clicks once, sees exactly what you can see, and can revoke control at any moment.
Multi-monitor & 4K
Switch displays without renegotiating the session. H.265 hardware encoding holds 60 fps at 4K, and drops to a low-bandwidth mode on bad hotel wifi.
Background tools
Terminal, PowerShell, file transfer, registry editor, service control, and task manager — without interrupting the person working on the machine.
Session recording
Every session is recorded and attached to the ticket, with keystroke and file-transfer logs. Retention is configurable up to 400 days.
In the ticket, not beside it
Sessions launch from the ticket you're already reading. Notes, recordings, and the run log land back on that ticket automatically — nothing to copy across.
Automation
The ticket is the trigger.
Every other RMM makes you write a script, attach it to a policy, schedule it, and hope the right machine matches. Lyvn reads the ticket in plain English, pulls the device's live state from the agent, and picks the runbook your team already wrote — then runs it on that endpoint and reports back on the ticket.
- Your scripts: PowerShell, Bash, Python, Zsh — imported, versioned, signed
- Your rules: auto-run, approval-gated, or maintenance-window only, per runbook
- Your rollback: snapshot before, verify after, revert and escalate if it fails
- Your receipts: full transcript and exit code attached to the ticket
Runbook library
42 activeStartup disk full
macos-disk-reclaim
Printer spooler stuck
win-spooler-restart
VPN client won't connect
vpn-reinstall
Outlook profile corrupted
outlook-profile-rebuild
Endpoint missing patches
patch-catchup
Suspicious process detected
isolate-endpoint
Signed · scoped · reversibleLyvn matched 1,204 tickets this month
Patching
Updates that don't start a fire.
Patch the fleet on a schedule you set, in rings, with the evidence your auditor wants already collected.
Third-party too
Chrome, Zoom, Slack, Acrobat, Java, and 300+ other apps — not just OS updates.
Ring deployment
Pilot ring gets the patch first. If the ticket volume stays flat for 48 hours, the rest of the fleet follows automatically.
Maintenance windows
Per-group schedules, reboot policies, deferral limits, and a hard override for emergency CVEs.
Compliance evidence
Per-device patch history exported for SOC 2, HIPAA, or cyber-insurance questionnaires.
Monitoring
Alerts that arrive as work, not noise.
Policies watch the signals that actually predict a ticket. When one fires, it opens a ticket with the device, the owner, and the suggested runbook already attached.
Disk above 90% for 2 hours
→ macos-disk-reclaim
SMART pre-fail on system drive
→ open P1, order replacement
Battery health under 80%
→ open asset renewal task
Backup agent stopped
→ restart, then escalate
Unexpected admin group change
→ page on-call, isolate
Outcomes
What happens when the endpoint fixes itself.
82%
Endpoint tickets closed by a runbook with no technician touch.
6.4m
Median time from alert firing to verified remediation.
↓ 91%
Drop in "can you remote in?" scheduling back-and-forth.
3→1
Tools replaced: RMM console, remote-desktop licence, patch tool.
vs.
Compared with a legacy RMM stack.
Capability
Legacy RMM
UniDesk
Remote desktop
Separate licence, separate console
Built in, launched from the ticket
Scripting
You write and schedule every script
Lyvn picks the runbook off the ticket
Alerts
Email storm into a console
Each alert becomes a triaged ticket
Ticketing
Integration, roughly synced
Same platform, same record
Agent footprint
Heavy, per-OS forks
One agent, ~40 MB, Win/macOS/Linux
Pricing
Per endpoint, per module
Flat platform fee, unlimited seats
Trust
An RMM is root on every machine you own. We treat it that way.
Remote-management tooling is one of the most attacked surfaces in IT, and letting an AI drive it only raises the stakes. So the guardrails aren't settings we bolted on — they're the default.
Scoped, short-lived credentials
Runbooks get a token for one device and one action, minted at run time and dead 60 seconds later. There is no standing domain-admin credential for an attacker to steal.
Signed scripts only
Every runbook is signed on publish. An unsigned or edited script will not execute on an endpoint, even if someone drops it into the library.
Approval gates you control
Auto-run is opt-in per runbook. Anything destructive — wipes, isolation, mass deployment — can require a named approver regardless of what the AI recommends.
Nothing happens off the record
Every session, keystroke, file transfer, and script run is logged against a ticket and a human owner. SOC 2 Type II, exportable to your SIEM.
Questions
Frequently asked.
Does UniDesk include remote desktop, or do I bring my own?+
It's built in. Attended and unattended remote control ship with the platform — no ScreenConnect, TeamViewer, or Splashtop licence to renew. Sessions start from the ticket, the device record, or a Slack message, and the recording lands back on the ticket when you disconnect.
Is RMM included in my plan?+
RMM is included on Growth and above. Starter covers ticketing, the guest portal, and asset inventory — endpoint monitoring, patching, and remote desktop unlock when you move up to Growth. There's no per-device fee at any tier: enroll the whole fleet and the plan covers it. Compare plans
How does the AI decide which script to run?+
Lyvn reads the ticket or alert, pulls the device's live state from the agent, and matches it against your runbook library — the same library your senior techs already maintain. Each runbook declares what it fixes, what it needs, and whether it may run unattended. If confidence is low or the runbook is gated, it drafts the plan and waits for a human to click execute.
Can I stop it from running things automatically?+
Yes, and that's the default for anything with teeth. Auto-run is opt-in per runbook. Most teams start with a handful of safe ones — disk cleanup, service restarts, cache clears — and widen the list once they've watched the run logs for a few weeks.
Which operating systems does the agent support?+
Windows 10/11 and Server 2016+, macOS 12+ on Intel and Apple silicon, and major Linux distributions (Ubuntu, Debian, RHEL, Rocky). One agent, roughly 40 MB, self-updating, with a hard CPU ceiling so it never becomes the performance ticket.
Can I write my own scripts?+
PowerShell, Bash, Python, and Zsh. Import what you already have, keep them in Git if you want, and set parameters, timeouts, expected exit codes, and rollback steps per runbook. Anything you can script, Lyvn can be told to run.
Does this replace my ticketing system too?+
It is the ticketing system — RMM and the helpdesk are the same platform, which is why a device alert and an employee complaint end up on the same record. See UniDesk Ticketing
Does it work for MSPs across multiple clients?+
Yes. Run unlimited client workspaces from one console with per-tenant runbooks, alert policies, and patch schedules. Technicians see a single queue across tenants; clients only ever see their own.
Start free
Deploy the agent to ten machines this afternoon.
One installer, one enrolment token, and you'll be remoting in and watching vitals before the coffee goes cold.
Start free trial →No charge today · 14 days · Unlimited technicians
See also
Ticketing →
The runbooks only fire because the ticket is smart enough to trigger them. See how intake, triage, and drafting work on the helpdesk side.
Explore ticketing ↗