# Extract a single frame as terminal output

**URL:** <https://discourse.asciinema.org/t/extract-a-single-frame-as-terminal-output/1047>\
**Category:** Ideas\
**Created:** [October 27, 2025, 2:38pm UTC](https://discourse.asciinema.org/t/extract-a-single-frame-as-terminal-output/1047 "2025-10-27T14:38:29Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![finite-state-machine](https://yyz1.discourse-cdn.com/flex027/user_avatar/discourse.asciinema.org/finite-state-machine/32/662_2.png) [@finite-state-machine](https://discourse.asciinema.org/u/finite-state-machine)\
**Post date:** [October 27, 2025, 2:38pm UTC](https://discourse.asciinema.org/t/extract-a-single-frame-as-terminal-output/1047/1 "2025-10-27T14:38:29Z")

</div>

As a user, I’d like to view or extract a single frame from a screencast at the command line. In other words, I’d like to “crush” the recording up to a certain point so that it can be written to the screen without writing to any cell more than once.

It looks like this functionality already exists for testing the avt/virtual terminal; it simply isn’t exposed at the command-line.

This would also facilitate manipulating recordings in a scriptable way, e.g., extracting a portion of the recording, removing a section from the middle, or greatly simplifying automated testing of other projects. (I haven’t come across any other terminal emulator that lends itself to running headlessly/without displaying anything.)

I’m guessing it wouldn’t be difficult to also provide the option to output a frame as svg, as another user has already requested.

Related requests:

> [@Single Image Output / Non-interactive "poster" mode](https://discourse.asciinema.org/t/single-image-output-non-interactive-poster-mode/915):
>
> Hey all wave I am a big fan of asciinema! Thanks for all of the hard work heart I am using a self-hosted asciicast for a documentation page I am working on. I like the new marker feature, but it has led me to the following question: I have a fairly long recording, and I would like to “expand” on it in the documentation. Is it possible to “only” show the image at that timestamp without having any interactive features? Ideally, I would like the single frame to be output to SVG, but since th…

> [@Export frame to html](https://discourse.asciinema.org/t/export-frame-to-html/968):
>
> Hello, Really enjoying asciinema. I wrote a [python script](https://github.com/DangerousPrototypes/BusPirate-docs-hugo/blob/main/tools/autodoc.py) that automates screencasts and marker placements to auto [update docs](https://docs.buspirate.com/docs/tutorial-basics/power-supply/) for a serial terminal device. The script writes the v2 asciinema format directly, and I use the player in hugo docs. Now I’m automating the static code snippets for the docs. An ansi2html script handles colors, but I’m eventually going to need more VT100 processing so everything looks right. It occurred to me that asciinema is doing something similar already. Is it po…