patat (Presentations Atop The ANSI Terminal) is a small
tool that allows you to show presentations using only an ANSI terminal. It does
- Leverages the great Pandoc library to support many input formats including Literate Haskell.
- Supports smart slide splitting.
- Slides can be split up into multiple fragments
- There is a live reload mode.
- Theming support.
- Auto advancing with configurable delay.
- Optionally re-wrapping text to terminal width with proper indentation.
- Syntax highlighting for nearly one hundred languages generated from Kate syntax files.
- Written in Haskell.
Table of Contents
- Table of Contents
- Input format
There is a pre-built package available for Debian:
Installation from source is very easy. You can build from source using
patat is also available from Hackage.
For people unfamiliar with the Haskell ecosystem, this means you can do either of the following:
- Install stack for your platform.
- Clone this repository.
stack setup(if you're running stack for the first time) and
- Make sure
$HOME/.local/binis in your
- Install cabal for your platform.
cabal install patat.
- Make sure
$HOME/.cabal/binis in your
patat [*options*] file
: If you provide the
patat will watch the presentation file
for changes and reload automatically. This is very useful when you are
writing the presentation.
: Run the presentation even if the terminal claims it does not support ANSI features.
: Just dump all the slides to stdout. This is useful for debugging.
: Display version information.
- Next slide:
- Previous slide:
- Go forward 10 slides:
- Go backward 10 slides:
- First slide:
- Last slide:
- Reload file:
r key is very useful since it allows you to preview your slides while you
are writing them. You can also use this to fix artifacts when the terminal is
The input format can be anything that Pandoc supports. Plain markdown is usually the most simple solution:
--- title: This is my presentation author: Jane Doe ... # This is a slide Slide contents. Yay. --- # Important title Things I like: - Markdown - Haskell - Pandoc
Horizontal rulers (
---) are used to split slides.
However, if you prefer not use these since they are a bit intrusive in the markdown, you can also start every slide with a header. In that case, the file should not contain a single horizontal ruler.
patat will pick the most deeply nested header (e.g.
h2) as the marker for a
new slide. Headers above the most deeply nested header (e.g.
h1) will turn
into title slides, which are displayed as as a slide containing only the
This means the following document is equivalent to the one we saw before:
--- title: This is my presentation author: Jane Doe ... # This is a slide Slide contents. Yay. # Important title Things I like: - Markdown - Haskell - Pandoc
And that following document contains three slides: a title slide, followed by two content slides.
--- title: This is my presentation author: Jane Doe ... # Chapter 1 ## This is a slide Slide contents. Yay. ## Another slide Things I like: - Markdown - Haskell - Pandoc
For more information, see Advanced slide splitting.
patat is fairly configurable. The configuration is done using YAML. There
are two places where you can put your configuration:
- In the presentation file itself, using the Pandoc metadata header.
For example, we set an option
val by using the following file:
--- title: Presentation with options author: John Doe patat: key: val ... Hello world.
Or we can use a normal presentation and have the following
Line wrapping can be enabled by setting
wrap: true in the configuration. This
will re-wrap all lines to fit the terminal width better.
autoAdvanceDelay to a number of seconds,
patat will automatically
advance to the next slide.
--- title: Auto-advance, yes please author: John Doe patat: autoAdvanceDelay: 2 ... Hello World! --- This slide will be shown two seconds after the presentation starts.
Note that changes to
autoAdvanceDelay are not picked up automatically if you
patat --watch. This requires restarting
Advanced slide splitting
You can control the way slide splitting works by setting the
variable. This variable defaults to the least header that occurs before a
non-header, but it can also be explicitly defined. For example, in the
following document, the
slideLevel defaults to 2:
# This is a slide ## This is a nested header This is some content
slideLevel 2, the
h1 will turn into a "title slide", and the
be displayed at the top of the second slide. We can customize this by setting
--- patat: slideLevel: 1 ... # This is a slide ## This is a nested header This is some content
Now, we will only see one slide, which contains a nested header.
By default, slides are always displayed "all at once". If you want to display them fragment by fragment, there are two ways to do that. The most common case is that lists should be displayed incrementally.
This can be configured by settings
true in the metadata
--- title: Presentation with incremental lists author: John Doe patat: incrementalLists: true ... - This list - is displayed - item by item
incrementalLists works on all lists in the presentation. To flip
the setting for a specific list, wrap it in a block quote. This will make the
list incremental if
incrementalLists is not set, and it will display the list
all at once if
incrementalLists is set to
This example contains a sublist which is also displayed incrementally, and then a sublist which is displayed all at once (by merit of the block quote).
--- title: Presentation with incremental lists author: John Doe patat: incrementalLists: true ... - This list - is displayed * item * by item - Or sometimes > * all at > * once
Another way to break up slides is to use a pagraph only containing three dots separated by spaces. For example, this slide has two pauses:
Legen . . . wait for it . . . Dary!
Colors and other properties can also be changed using this configuration. For example, we can have:
--- author: 'Jasper Van der Jeugt' title: 'This is a test' patat: wrap: true theme: emph: [vividBlue, onVividBlack, bold] imageTarget: [onDullWhite, vividRed] ... # This is a presentation This is _emph_ text. ![Hello](foo.png)
The properties that can be given a list of styles are:
The accepted styles are:
As part of theming, syntax highlighting is also configurable. This can be configured like this:
--- patat: theme: syntaxHighlighting: decVal: [bold, onDullRed] ... ...
decVal refers to "decimal values". This is known as a "token type". For a
full list of token types, see this list -- the names are derived from there in
an obvious way.
"Patat" is the Flemish word for a simple potato. Dutch people also use it to refer to French Fries but I don't really do that -- in Belgium we just call fries "Frieten".
The idea of
patat is largely based upon MDP which is in turn based upon
VTMC. I wanted to write a clone using Pandoc because I ran into a markdown
parsing bug in MDP which I could not work around. A second reason to do a
Pandoc-based tool was that I would be able to use Literate Haskell as well.
Lastly, I also prefer not to install Node.js on my machine if I can avoid it.