Product

Remote terminal: a live shell from the browser

The remote terminal opens a live shell session against a managed device, from the browser and with nobody on the other side. Output arrives live, the keyboard is connected, and Ctrl+C genuinely interrupts the process that is running.

It is a real shell, not a command box

Today it runs on Windows devices, where the agent opens PowerShell in the SYSTEM context. That is the difference from the AI command console: there you send a command and wait for the result; here there is a live process on the other side to hold a conversation with.

A program asking for a password can be answered without what you type being visible. And for the program that asks without turning off the echo, the technician switches on the output hiding mode by hand while typing it.

What was seen is recorded, what was typed never is

The session is recorded, and the rule is strict: the output is recorded, meaning what appeared on screen, and never what the technician types. Of the typing, only when and how many characters are kept, never the content. The output already echoes back every key when echo is on, and it does not echo it back precisely when someone is typing a password.

Recordings are retained for 90 days as a floor, and that floor cannot be lowered. Retention can be raised, never lowered. The reason is not to protect the user: an MSP that can set retention to zero turns its own audit trail into decoration.

The limits are set, and explained

There are three simultaneous sessions per organization and only one per device. The per-device limit is not negotiable upward: two technicians typing into the same shell step on each other's keys and the resulting recording is unreadable.

The session closes itself after 15 minutes without activity, and has a hard cap of 4 hours even while in use. A forgotten shell left open against a customer's machine is an open door. If the agent goes quiet for 60 seconds, the session is treated as lost: a terminal that has been mute for a minute is no longer usable for work.

Who can open one and who can read it

Opening a terminal requires its own permission on that device, separate from the one for running commands. Reading the recordings requires the device management permission: being able to open a shell against one machine and being able to read everything everyone else did are two different things.

Frequently asked questions

Which systems does the remote terminal work on?

On Windows devices, where the agent opens PowerShell in the SYSTEM context. On a device whose platform is not implemented, the panel shows the button switched off instead of opening a session that will not answer.

Can I interrupt a process with Ctrl+C?

Yes, it is a real Ctrl+C: the agent translates it into the interrupt byte on the console input, just as if you were sitting at the machine.

Is what I type recorded?

No. The output is recorded, meaning what appears on screen. Of what the technician types, only when and how many characters are kept, never the content. And for a password prompt that does not turn off the echo, there is the output hiding mode.

How long are recordings kept?

90 days minimum. Retention can be raised per organization, never lowered below that floor: the code raises it back anyway.

How many terminals can I have open at once?

Three per organization and one per device. The per-device cap exists because two technicians on the same shell step on each other's keys.

Can a session stay open forever?

No. It closes itself after 15 minutes without activity and has a hard cap of 4 hours. The agent also stops on its own when the session permit expires, so a session cannot stay alive because the panel went away.

Try KairosLink free for 14 days

All modules included. No credit card.