Dialogue Assembler

Cuts speaker clips into single lines using the script, puts them in order and adds the transitions. Everything runs in your browser. Your files are never uploaded.

1. Add the files

Drag the whole folder here (or all files at once)

Choose folder Choose files

How to name files and tag the script

One .txt script. Dialogue lines look like Mentor: text. Titles and other text are ignored.
Speaker clips have the speaker name in the filename: mentor_1.mp4, husband_2.mp4.
transition.mp3 is the whoosh for every slide.
Tags in the script, on their own line or before a line:

[intro.mp4]                          whole clip, muted
[left]                               slide-left + whoosh into the next item
[background.mp4 + background.mp3]    whole clip, muted, with the mp3
[background.mp4 + background.mp3 @1.2]   mp3 starts 1.2 s into the clip
[left]
Mentor: So, what happened with the doctor?
Husband: You were right. ...

Settings

Whisper model Slide duration (s) Whoosh volume (1 = original) Default mp3 start after clip start (s) mp3 volume Padding around each line (s) Silence threshold (dB, -30 noisy / -40 quiet) Minimum match (%) Output quality (CRF, lower = better) Encoding speed

2. Analyze

Transcribes the clips and finds every script line. The first run downloads the Whisper model; later runs reuse it and the saved transcripts.

Add files to start.

Technical log