I've made a version of it you can listen to: https://pdftomp3.com/player.html?example_id=6ad14c5229ba4da7...

I've updated my video to show the player experience and the quality of the voices:

https://youtu.be/1qiH_SeDyyc

Here you go my friend, a two speaker podcast in Thai :)

https://pdftomp3.com/published/player.html?publication_id=6a...

Its about the whole pipeline. From adding a PDF to listening in the car (with the app).

It handles the formatting of scientific papers very well. E.g. I can exclude footnotes etc. But I didn't test it much with tables yet.

I went the route of handling epub first (really, anything pandoc can convert to epub) on ebookaloud, and it took quite a few iterations to get consistent parsing across a wide corpus of test documents.

My experience is that gathering and testing with that corpus was about 30% of the overall effort. I have not yet gathered the motivation to tackle pdf: it’s more of a layout format than a text format. Not just tables but multi-column layouts and similar oddities leave a pile of imperfect heuristics in their wake.

Multi-lingual support is even worse. I was working with arabic script in pdf years ago and they don’t even capture the encoded text, just rendering instructions due to the high number of ligatures. I don’t know if things have improved since then.

makes sense, tables in pdfs are a nightmare anyway. cool project.

Here you can listen to examples: https://pdftomp3.com/examples/

Via MCP more voices are available (all OpenAI and Gemini voices, depending on which api key you use).

I tested it - the word highlight and audio spoken did not always align, leading to a discordant experience.

The experience has been improved now!

I like the voices, will be using the product soon. Thanks for sharing!

Thanks! Currently, I'm using Gemini TTS and Kokoro TTS as a free option. Kokoro is fast and I can run it on my own server, but I'm not using it as a standard yet because Gemini supports more languages and has more voices available. But text-to-speech processing is the biggest bottleneck at the moment.

I did both of that too! At the end I made my own app to listen to the episodes.

There are two parts to the processing. The processing of the document itself is done on your OpenAI/Codex subscription. Additionally, you have to submit an API key for the text-to-speech processing. There you can use a Gemini key or an OpenAI key.