Tutorial

Vera uploads a transcript

Vera is a paralegal. In this tutorial she uploads a hearing, picks a voice for each speaker, and listens with the words in front of her.

Poster for the tutorial. A folder and headphones, marked review draft. The video is not included.

Review draft This tutorial video is not on the site yet.

Sample coming soon.

The picture is a poster, not the video. The video is still a review draft, so it is not included. The hearing below is invented. Ada Okonkwo, Ellis Vell, Rowan Pell, and Example County Court are not real.

Read along

Press play and the current line and speaker light up. Click a line, or a word, and the audio jumps to that moment. The times are the same clock as the WebVTT file and the sync JSON the app writes. A line here is one spoken sentence, not a printed page-and-line number.

Sample coming soon.

When the sample recording is ready, a short Docket Brief plays first. It is labeled as a summary and is not part of the verbatim record. The hearing follows it. Try the voice choices for these speakers.

  1. THE COURT

    Good morning. We are here in Example County Court on State versus Pell. Please state your appearances.

  2. MS. OKONKWO

    Ada Okonkwo for the prosecution, Your Honor.

  3. MR. VELL

    Ellis Vell for the defense.

  4. MS. OKONKWO

    The State calls Rowan Pell. Did you write the notebook marked Exhibit A?

  5. THE WITNESS

    I did. I wrote those pages in March, at my kitchen table in Example County.

  6. MR. VELL

    Objection. That notebook was not given to the defense before today.

  7. THE WITNESS

    I kept the notebook in a blue box. I never gave it to the prosecution.

  8. THE COURT

    The objection is noted. I will rule on the notebook after the noon recess. The witness is excused until two o'clock.

Words on screen

The MP4 shows the speaker's name and the current line while the recording plays. The picture lights the whole line, not one word at a time. Word times stay in the sync file, which is what the read-along player above uses when you click a word.

Sample coming soon.

The words-on-screen video will play here once the sample file is in place.

What you get

A render writes these files next to each other. The invite upload today gives you the MP3, the SRT and VTT caption files, and the private listen page. The MP4 is written when you ask for the video. Nothing else is invented here.

  • MP3 audio

    .mp3 The recording. ID3 tags hold the title, case, court, and date. Chapter marks are included when chapters are on. A Docket Brief, when it is on, is spoken at the start of this file. It is not a separate download.

  • Read-along video

    .mp4 The speaker's name and the current line, on a plain screen. Written when you ask for the video. Building it needs ffmpeg.

  • SRT captions

    .srt One cue per spoken sentence, with the speaker label, timed to the merged audio. Includes the Docket Brief when that summary was spoken.

  • WebVTT captions

    .vtt The same cues as a WebVTT file, for players that use that format.

  • Click-sync map

    .sync.json Line and word times in milliseconds from the start of the MP3. The read-along player uses this file when it is present, and falls back to the WebVTT line times.

  • Listen-along page

    .sync.html A page that sits next to the MP3. Click a line or a word and the audio jumps to that moment.

  • Index card

    .summary.txt The case, the court, the date, and the conventional file name, in a small text file.

  • Spoken-accuracy report

    .accuracy.txt and .accuracy.json When the check runs, the spots that still do not match the page, with the time to jump to. A silent dry run does not write them.