Videos can be very effective in teaching, but I myself never make them because I don't like narrating videos nor video editing. So that got me thinking whether text-to-speech can be used to generate videos automatically. That's what I hacked together now in the `trv` Rust crate (it's a binary that you can install via cargo install, see https://github.com/transformrs/trv for the source code and docs).

It's a tool that you can give a Typst presentation with speaker notes to. Next, the tool will turn the Typst file into images and audio, and then turn everything into a video.

For example, I made one I video about a blog post that I wrote earlier. Unfortunately, I cannot directly upload a video here on HN, so here is a link: https://youtu.be/vn8-Asioxq8.

To give an idea of how the video was made, here is the first slide of the Typst presentation:

    #import "@preview/polylux:0.4.0": *

    #set page(paper: "presentation-16-9", margin: 1in)
    #set text(size: 30pt)

    #slide[
        #toolbox.pdfpc.speaker-note(
        ```md
        Iterators are pretty cool.
    
        For example, in Python we could write the following code in a normal loop.

        Here we have a list of 3 values and we add 1 to each value.

        This returns a new list with the values `[2, 3, 4]`.
        ```
        )

        ```python
        values = [1, 2, 3]

        for i in range(len(values)):
            values[i] += 1

        print(values)
        # [2, 3, 4]
        ```
    ]

Next, I ran the following command:

    $ trv --input=presentation.typ \
          --model='hexgrad/Kokoro-82M' \
          --voice='am_liam' \
          --release"

This created a video of 1.2 MB that I then uploaded to YouTube.

Is a tool like this useful? What are your thoughts?

Submissions by huijzer

Show HN: trv. A tool to turn a presentation and speaker notes into a video

Show HN: trv. A tool to turn a presentation and speaker notes into a video