Research papers, in your podcast app

The paper you've had open in a tab all week can come with you on a walk. Paste its arXiv link, or upload the PDF, and Askwho Casts Pro reads it into your own private podcast feed, equations, figures and footnotes included.

Try your free article

Your first article is free, up to 30 minutes. No card needed.

Hear a real one

Real Askwho Casts audio · played at 1× The Platonic Representation Hypothesis Minyoung Huh, Brian Cheung, Tongzhou Wang and Phillip Isola

The cast, in order of appearance

  1. Image descriptionsvoiced by Zinnia Voss
  2. Narratorvoiced by Laurence Ford
  3. Footnotesvoiced by Corinne Langley

From the introduction: Figure 1 described, the paragraph that refers to it, and its footnote. The whole episode runs 53 minutes.

Read it at arXiv 2405.07987. Used under CC BY 4.0. This is a reading of its main text, with its figures described aloud. Its author affiliations, acknowledgements, appendices and reference list are left out.

From a link or a file

  • An arXiv link

    Paste the paper's arXiv address into your Library. The abstract page, the PDF link and the web version all work. Askwho Casts Pro reads arXiv's web version of the paper, which most papers since late 2023 have, and many older ones too.

  • A PDF

    If a paper has no web version, or it came from a journal, upload the PDF from your Library, up to 80 MB. A link to a PDF can't be read, so it's the file itself that's needed. Uploads come with a plan.

  • Any open web page

    A paper or preprint published as an ordinary web page can be pasted like any article.

Whichever way it comes in, you see the running time and the price before anything is made.

What you'll hear

The paper's own words, not a summary. Its main text is read from the abstract to the conclusion, each section under its own heading, and the headings become chapters in your podcast app.

From a link, equations are read aloud where they stand in the sentence, rather than skipped. In an uploaded PDF most are too, and one that can't be read reliably is announced as an equation, with its number, rather than garbled. Figures are described in one to three sentences each, alongside their captions, so you know what the chart showed.

A footnote is read just after the sentence it belongs to, and a footnote that is nothing but a link isn't read. From a link, the reference list, the acknowledgements and the appendices are left out of the audio, and the show notes link back to the paper, where they are. With an uploaded PDF you choose, as below.

When you upload the PDF

An uploaded paper is checked over before it's made. You see the abstract and the sections it found, and the parts that don't read well aloud start switched off: the title page's affiliations, the acknowledgements, the appendices and the references. Switch any of them back on if you'd like to hear them.

Bracketed citation numbers are dropped, and citations by name, such as "Smith et al., 2020", are read the way people say them aloud. Page numbers and running headers are left out. The show notes carry the authors, the year and the paper's DOI or arXiv number.

What it costs

One credit is about one minute of finished audio, and you see the exact price before you make anything. An uploaded PDF costs 1.25 credits a minute of finished audio.

Plans start at $9 a month for 200 credits.

Where it plays

Your feed has its own private address. Add it to Apple Podcasts, Overcast, Pocket Casts, AntennaPod, Podcast Addict and most other podcast apps. Spotify and YouTube Music can't add private feeds.

Every episode also plays on the site, with a read-along that lights each line as it's spoken.

Start with a paper you meant to read

Your first article is free, up to 30 minutes. No card needed.

Try your free article

After that, plans start at $9 a month for 200 credits. One credit is about one minute of finished audio, and you see the exact price before you make anything.