Obsidian SuiteDocumentation
Obsidian Remote: all chapters

Docs / Obsidian Remote / Connecting

Connecting to a computer

To reach another PC you need its ID (or its network address) and its permission: its password, or the person at it clicking Accept.

Connect by ID

  1. Ask for the remote PC's 9-digit ID. It is at the top of that PC's main window, under Your ID.
  2. Type it in the Control a remote desktop box. Spaces and dashes don't matter: 123 456 789, 123-456-789 and 123456789 are the same.
  3. Press Connect, or one of the buttons below the box for another kind of session.
  4. Sign in to the remote PC as it asks (see Signing in to the remote PC).

Connecting by ID needs a server: if none is set up, the app says No server is set up for IDs - add one in Settings > Network, or connect by IP address. You can't connect to your own ID (That is this PC's own ID.).

You can also connect from the device lists: double-click a device, or right-click it and pick the kind of session. See The main window.

The kinds of session

ButtonWhat you get
ConnectFull control: the picture, mouse, keyboard, clipboard, sound, files, terminal and tunnels, as far as the remote PC allows.
View onlyThe picture and the sound. You can't use the mouse or keyboard, the clipboard or files.
Transfer filesThe file manager, without the picture. See Transferring files.
TerminalA PowerShell window on the remote PC, without the picture. See Remote terminal.
TunnelPort forwarding, for example for Remote Desktop, without the picture. See TCP tunnels.

The remote PC only grants what the kind of session needs: a view-only session can never type, even if the remote PC would allow it in a full session.

How the connection travels

When you connect by ID, the server finds the remote PC and vouches for its key. Then the app tries two routes at the same time and keeps whichever works first:

  • Direct (LAN): straight to the remote PC, when both PCs are on the same network (or at the same site). This is the fastest.
  • Relay: through your server, when the two PCs can't reach each other directly. Still end-to-end encrypted; the server only passes the encrypted traffic along.

The session window's status line shows which one you got. See Local network and Wake-on-LAN for what makes direct connections possible.

Connect by address

You can also connect straight to a PC's network address, without a server, for example on a site without one:

  • Type the address in the box: 10.0.0.20, 10.0.0.20:21118, a name like pc1.example.com, or an IPv6 address in brackets like [fd00::20]. Without a port, the app uses 21118.
  • The remote PC must allow it: Allow direct IP access without the server in its Settings > Security (it is off at first). Otherwise you see The device closed the connection - direct IP access may be off there.
  • Your PC must be able to reach the remote PC's port 21118, which normally means the same network.

Because no server vouches for a PC you reach by address, the app shows its key the first time (New device). Ask the person at the remote PC to read you the key from their Settings > About (or the Key line in their main window), compare, and click Trust and connect. The app remembers the key.

When a device's key changes

Every Obsidian Remote PC has its own key, and the app remembers the key of each PC you connect to. If a PC turns up with a different key, the app stops and shows Device key changed, with the old and the new key.

That happens when Obsidian Remote was reinstalled on that PC with its data removed, or when the PC was replaced. It is also what it would look like if someone were sitting between you and the PC. Only click Connect anyway if you know why the key changed; if in doubt, ask the person at the PC to read you the key from their Settings > About.

Signing in to the remote PC

What happens next depends on how the remote PC accepts connections (set in its Settings > Security, or by your administrator's policy):

The remote PC acceptsWhat the app does
Password only (unattended)It asks for the Password. Type the one-time password shown in that PC's main window, or its permanent password, and click Connect.
Only when the person at this PC clicks AcceptIt shows Waiting for the remote user to accept... while the person there sees your request. They have 60 seconds to answer.
Password, or the person at this PC clicks Accept (the usual setting)It asks for the Password. Type it and click Connect, or click Ask them to accept (or Connect with the box empty) and it waits for the person there, as above. If you have the device in your personal address book with a saved password, it uses that without asking.

If the person there clicks Deny, or nobody answers within 60 seconds, a PC that also takes a password asks you for it: The remote user refused the connection. You can use the password instead. (or Nobody accepted the connection in time.). The person there is asked only once per connection. The web viewer works the same way (see Connecting from a browser).

If the password is wrong, the app says Wrong password. and asks again. The one-time password changes after every session that used it, so make sure you have the current one.

If the remote PC uses two-factor sign-in, the app then asks for a Two-factor code: the 6-digit code from the authenticator app that was set up for that PC. Leave Don't ask again on this PC ticked and it won't ask on this PC again for 90 days. See Two-factor sign-in.

To protect the remote PC from guessing, it limits tries:

  • 6 wrong passwords within a minute from one address lock that address out for a minute.
  • 30 wrong passwords within 10 minutes from anywhere stop the PC accepting passwords for 5 minutes. Accept still works then.
  • The whole sign-in, including waiting for Accept, must be done within 150 seconds.

A link like obsidian-remote://connect/123456789 opens the app and connects to that ID. Add ?kind=view, ?kind=file, ?kind=terminal or ?kind=tunnel for another kind of session. Links only accept IDs, never addresses, so a web page can't point your app at an arbitrary computer, and the app asks Open a link? with the ID (and the computer's name, if you connected to it before) before it connects.

If it doesn't connect

The session window (or the small connecting window, for files, terminal and tunnels) says what went wrong, with Try again and Close. The most common messages:

  • The device is offline. The remote PC is off, asleep or not connected to the server. Try waking it (see Local network and Wake-on-LAN).
  • No device has that ID on this server. Check the ID, and that both PCs use the same server.
  • The connection was refused (by the remote user, or the server does not allow it). The person clicked Deny (on a PC that takes only Accept, or you gave up on the password), or you don't have access to that device.
  • No answer in time (from the device or the remote user). Nobody clicked Accept within 60 seconds, or the PC didn't answer.

Troubleshooting lists every message and what to do about it.