#snapshot #terminal #svg

app term-transcript-cli

CLI wrapper for term-transcript

6 releases

0.3.0 Jun 3, 2023
0.3.0-beta.2 Apr 29, 2023
0.3.0-beta.1 Jan 19, 2023
0.2.0 Jun 12, 2022
0.1.0 Jun 1, 2021

#5 in #snapshot



term-transcript CLI

Build Status License: MIT OR Apache-2.0 rust 1.66+ required

This crate provides command-line interface for term-transcript. It allows capturing terminal output to SVG and testing the captured snapshots.


Install with

cargo install --locked term-transcript-cli
# This will install `term-transcript` executable, which can be checked
# as follows:
term-transcript --help

Alternatively, you may use the app Docker image as described below, or download a pre-built app binary for popular targets (x86_64 for Linux / macOS / Windows and AArch64 for macOS) from GitHub Releases.

Crate feature: portable-pty

Specify --features portable-pty in the installation command to enable the pseudo-terminal (PTY) support (note that PTY capturing still needs to be explicitly switched on when running term-transcript commands). Without this feature, console app output is captured via OS pipes, which means that programs dependent on isatty checks or getting term size can produce different output than if launched in an actual shell (no coloring, no line wrapping etc.).

Crate feature: tracing

Specify --features tracing in the installation command to enable tracing of the main performed operations. This could be useful for debugging purposes. Tracing is performed with the term_transcript::* targets, mostly on the DEBUG level. Tracing events are output to the stderr using the standard subscriber; its filtering can be configured using the RUST_LOG env variable (e.g., RUST_LOG=term_transcript=debug).


  • The capture subcommand captures output from stdin, renders it to SVG and outputs SVG to stdout.
  • The exec subcommand executes one or more commands in the shell, captures their outputs, renders to an SVG image and outputs it to stdout.
  • The test subcommand allows testing snapshots from the command line.
  • The print subcommand parses an SVG snapshot and outputs it to the command line.

Launch the CLI app with the --help option for more details about arguments for each subcommand. See also the FAQ for some tips and troubleshooting advice.

Using Docker image

As a lower-cost alternative to the local installation, you may install and use the CLI app from the GitHub Container registry. To run the app in a Docker container, use a command like

docker run -i --rm --env COLOR=always \
  ghcr.io/slowli/term-transcript:master \
  print - < examples/rainbow.svg

Here, the COLOR env variable sets the coloring preference for the output, and the - arg for the print subcommand instructs reading from stdin.

Running exec and test subcommands from a Docker container is more tricky since normally this would require taking the entire environment for the executed commands into the container. In order to avoid this, you can establish a bidirectional channel with the host using nc, which is pre-installed in the Docker image:

docker run --rm -v /tmp/shell.sock:/tmp/shell.sock \
  ghcr.io/slowli/term-transcript:master \
  exec --shell nc --echoing --args=-U --args=/tmp/shell.sock 'ls -al'

Here, the complete shell command connects nc to the Unix domain socket at /tmp/shell.sock, which is mounted to the container using the -v option.

On the host side, connecting the bash shell to the socket could look like this:

mkfifo /tmp/shell.fifo
cat /tmp/shell.fifo | bash -i 2>&1 | nc -lU /tmp/shell.sock > /tmp/shell.fifo &

Here, /tmp/shell.fifo is a FIFO pipe used to exchange data between nc and bash. The drawback of this approach is that the shell executable would not run in a (pseudo-)terminal and thus could look differently (no coloring etc.). To connect a shell in a pseudo-terminal, you can use socat, changing the host command as follows:

socat UNIX-LISTEN:/tmp/shell.sock,fork EXEC:"bash -i",pty,setsid,ctty,stderr &

TCP sockets can be used instead of Unix sockets, but are not recommended if Unix sockets are available since they are less secure. Indeed, care should be taken that the host "server" is not bound to a publicly accessible IP address, which would create a remote execution backdoor to the host system. As usual, caveats apply; e.g., one can spawn the shell in another Docker container connecting it and the term-transcript container in a single Docker network. In this case, TCP sockets are secure and arguably easier to use given Docker built-in DNS resolution machinery.


This example creates a snapshot of the rainbow script and then tests it.

Testing rainbow example

The snapshot itself is tested, too! It also shows that SVG output by the program is editable; in the snapshot, this is used to highlight command-line args and to change color of comments in the user inputs.

The test command can compare colors as well:

Testing color match

Another snapshot created by capturing help output from a pseudo-terminal (the --pty flag):

Output of test-transcript --help

Using PTY enables coloring output by default and formatting dependent on the terminal size.

See also a shell script used in the "parent" term-transcript crate to render examples; it uses all major commands and options of the CLI app. The snapshots generated by the script are showcased in a dedicated file.


Licensed under either of Apache License, Version 2.0 or MIT license at your option.

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in term-transcript by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.


~202K SLoC