Can it be done without installing a client on both ends

Tidying the icons on an older relative's desktop, looking at the error on a colleague's machine, reaching the desktop at home while travelling: none of this happens often, and every time it runs into the same obstacle. Remote tools generally want a client on both machines, and the other machine is usually one where installing software is not allowed. Chrome Remote Desktop splits the problem into two unequal halves. The machine being controlled installs a host component once. The person connecting needs only a browser, signed into the same account, and sees a list of their own devices on a web page. The tool is maintained by Google and has been around since 2011; as of September 2026 the host component stands at 152.0.7977.9.

Two ways to use it: your own machines, or a one time favour

The entrance is divided in two. Remote access is the standing arrangement: the host component is installed on the machine to be controlled, given a name and a PIN, after which that machine sits under the account. From any device, open a browser, enter the PIN, and the session opens; the phone and tablet apps connect in the same way. Remote support is the one time version: whoever shares their screen generates an access code and hands it over, the code stops working once used, and during a session a confirmation prompt appears from time to time so that an authorisation is not left hanging for a long while. Both modes use the same channel underneath, and differ only in how the permission is granted: a fixed device under one's own account, or a door opened for a short while.

A short list of features, with the useful ones present

The clipboard synchronises in both directions, so a block of text can be copied on one side and pasted on the other without a detour through cloud storage. Audio is forwarded as well, which means sound playing on the distant machine arrives at the connecting end, so a video or a voice message does not have to be guessed at from the picture. A machine with several monitors can show all of them and switch between them one at a time, and system key combinations and screen capture keys are sent from a side menu; when both ends have a physical keyboard, keys can be mapped as well. Register the host component as a system service and nobody has to be sitting at the distant machine to approve a connection: it is reachable as soon as it is switched on and online.

What it deliberately does not do

The purpose of the tool is to put a screen in front of you, and anything that would need a second channel is largely absent. There is no file transfer panel inside a session and no remote printing. No recording is kept afterwards and the history of access does not become an exportable record. Moving a file means using cloud storage or mail, in other words a separate path. The picture refreshes when it needs to, which makes document editing, browsing and changing settings feel close to the machine itself, but work that demands a high frame rate and changes the whole screen constantly was not what this was designed for.

The other condition is that the host must be switched on and connected. A sleeping machine will not wake itself because somebody asked to connect. Cross system use does work: the host component installs on Windows, macOS and Linux, and the connecting side can be a browser on any of those or a phone. Resolution and scaling follow the actual window, so a narrow screen means a certain amount of scrolling.

The account is the permission, which sets the boundary

Opening a connection requires two credentials to hold at once: the identity of the account and the machine's own PIN. The channel is encrypted, and the PIN is set and changed locally by whoever uses the machine, so it does not have to be remembered anywhere else. This also means the security boundary is roughly the security of the account itself. Two step verification on the account raises the bar for remote access by the same amount; conversely, a compromised account exposes every device listed under it. The device list, resetting a PIN and starting a session all happen on one web page, with no second administration console to maintain.

If the only need is to show a screen to somebody for a short while, a code that expires after use is safer than a permission left open for a long time. The confirmation that keeps appearing during a shared session serves the same purpose: it reminds the person sharing that this door is still open.

Who it suits, and who it does not

It fits personal and household use: moving between one's own machines, keeping an eye on an older relative's computer, reaching a powerful desktop from a thin laptop, or connecting briefly to one's own machine from a public computer where nothing may be installed. It does not fit the case of managing dozens of devices, handing out permissions by role and keeping an audit trail of who connected when; there is no such capability here, and another arrangement is required. On price, personal use costs nothing and there is no paid tier; the cost amounts to needing an account.

The trade is plain: installing once on the host and nothing at all on the connecting side buys an extremely low cost of entry, and the price paid is a deliberately restrained list of features. No file transfer, no audit trail. For personal use and the occasional favour of looking at somebody's screen, it is exactly as much trouble as the job deserves, and no more.