Podcast chapter timestamps from a transcript

All three read the same input: an SRT or VTT file, or text that has a time at the start of each line, such as the timestamped TXT export ([05:12] Maria: ...). Without times there is nothing to place chapters or quotes against, and the page says so. A Name: label at the start of a line is read as the speaker.

How the chapter points are chosen

The first chapter is always at 0:00. For each later one the tool looks at the lines that start within a quarter of the spacing either side of the ideal time (for example 4 to 6 minutes for a chapter at about 5) and takes, in this order:

1. the line that follows the longest pause, when the file has end times (an SRT or VTT does; a pause of at least 0.7 seconds counts);
2. otherwise a line that follows a full stop;
3. otherwise the nearest line.

Pauses and full stops are hints, not proof. A speaker often pauses before a new topic, and also before a joke. The page tells you which rule was used for your file so you know how much to trust the placement. Titles are placeholders: the three words most particular to that stretch (words that are common across the whole episode are ignored), the first words of the line, or a number.

Three forms of the same list

• A list for the description, one chapter per line (12:30 Pricing), which is how YouTube reads chapters and how many show-notes pages are written. The tool checks it against YouTube's rules for chapters and tells you if it fails: the first stamp 0:00, at least three, ascending, ten seconds minimum (a short episode may not reach three).
• A Podcasting 2.0 chapters file (JSON, version 1.2.0 as in the Podcast Index's example): each chapter has a startTime in seconds and a title. The spec says the file should be served with the content type application/json+chapters and referenced by <podcast:chapters url="..." type="application/json+chapters"/> in the episode's feed item. That is something your host or feed has to do.
• A WebVTT chapters file, one cue per chapter with its start and end, the kind an HTML <track kind="chapters"> element can load for a web player.

{
  "version": "1.2.0",
  "chapters": [
    { "startTime": 0, "title": "Intro" },
    { "startTime": 168, "title": "Pricing" }
  ]
}

(Shape shown for illustration; the page fills in your own times and titles.)

What it does not do

It does not name chapters well, add chapter images or links, write the chapter data into an MP3 file, or publish anything. A podcast editor can embed chapters in the audio; this page writes the plain list and the files. If your transcript has no times it stops and says so. The TryScriptorium app's own Markdown export has a Chapters section, split at the longest silences and labelled with distinctive words, when a transcript has more than one; this page differs in that you choose the spacing and the output form, and in that it works on any transcript you paste, not only ones made here.

Check before you publish

Play the audio at a few of the chapter times. A chapter that starts in the middle of an answer is better moved to the start of the question, which means editing the time by hand. Then paste the list into the YouTube chapters checker if it is going into a video description. If the text came from YouTube's transcript panel, clean it first with the YouTube transcript cleaner, which can also draft chapters from that text.