Seamless mode

August 16, 2026 · View on GitHub

Seamless mode is Xpra’s usual mode for forwarding individual applications. It is started with xpra seamless (or the legacy alias xpra start). Each remote window appears alongside the client’s local applications instead of inside a single remote-desktop window.

Why use seamless mode?

Native window management

The client’s operating system and window manager control each forwarded window. Moving, minimizing, resizing, and switching windows therefore remain responsive even when the connection has noticeable latency.

No remote desktop canvas

Unlike desktop, monitor, and shadow mode, seamless mode does not stream a complete desktop or screen. There is no large remote desktop window to navigate, and applications can be placed naturally across the client’s monitors.

Efficient, focused sessions

Only the windows and features you choose to forward are presented to the client. This is usually easier to use and wastes less bandwidth than encoding an entire desktop, especially when only a few applications are needed.

With the HTML5 client, forwarded windows remain inside the browser’s canvas; native desktop placement requires a regular Xpra client.

Display backends

The X11 backend is the default:

xpra seamless --backend=x11 --start=xterm

The newer Wayland backend can be selected explicitly:

xpra seamless --backend=wayland --start=weston-terminal

The Wayland backend is experimental and may require the separate xpra-server-wayland package. Use the X11 backend when you need the broadest compatibility.

Limitations

Seamless server sessions require a supported display environment and are not available on Windows or macOS servers. On those platforms, use shadow mode to access an existing desktop or choose a different session mode.

Some experiments attempt to shadow only selected applications or windows as an approximation of seamless mode, but this is not reliable enough to recommend as a general solution. See issue #3476 for the current status.