Pseudoterm

Pseudoterm, short for pseudo-terminal, is a software abstraction that provides a terminal interface to a program running in an environment that is not itself a real terminal. It enables terminal-like interaction by pairing a master endpoint with a slave endpoint; the master is used by a terminal emulator or controlling process to send input and read output, while the slave presents a terminal interface to the running application as if it were connected to a physical terminal.

In Unix-like systems, the kernel implements a PTY (pseudo-terminal) subsystem. The master device is typically accessed through a multiplexer such as /dev/ptmx, which creates a new slave device under /dev/pts/N. The slave side is used by the application as its controlling terminal. Typical workflow includes allocating a pty with posix_openpt, enabling it with grantpt and unlockpt, and locating the corresponding slave path with ptsname. The child process may then create a new session, set the slave as its controlling terminal, and duplicate the slave descriptor to standard input, output, and error. The terminal settings (via termios) and line discipline apply to the slave side.

Pseudoterms are widely used to enable interactive sessions in environments where a real terminal is not available, such as remote shells, terminal multiplexers, automated testing, and scripting frameworks. They also underpin tools like SSH, which allocates a pty for remote sessions, and terminal multiplexers such as screen or tmux, which run multiple pseudo-terminals under a single user interface.

Security and compatibility considerations include proper permissions on the slave devices, correct handling of controlling terminals, and ensuring that pty usage does not bypass access controls. Windows provides a related concept called the pseudo console (ConPTY) for similar functionality in that environment. See also PTY, termios, SSH, screen, and tmux.