The transcript is analysed for topic shifts and long pauses. Chapter markers with descriptive titles are placed at every natural boundary, exportable to YouTube chapters, Spotify chapters and podcast show notes.
The chapter detection algorithm analyses the transcript using a sliding window approach. It measures the semantic similarity between adjacent transcript segments using keyword overlap and topic vocabulary. When the similarity drops below a threshold — indicating a genuine topic change — a chapter boundary is placed at that point.
Chapter boundaries are also confirmed by long pauses in the audio, since speakers naturally pause when transitioning between major topics. The two signals — semantic shift and audio pause — are combined to produce boundary candidates with a confidence score.
The minimum chapter duration setting prevents chapters from being too short. YouTube recommends chapters be at least 10 seconds long to appear in the chapter list. Podcast platforms typically work better with chapters spaced 3–10 minutes apart. EditLint AI respects your minimum spacing setting when selecting which boundaries to keep.
After placing chapter markers on your timeline, export the chapter list in any of the supported formats from the EditLint panel.
0:00 Introduction)Local Whisper transcription produces the full transcript. The transcript is split into segments based on natural sentence boundaries. Each segment is represented as a bag of keywords for semantic similarity comparison.
Adjacent segments are compared for semantic similarity. Boundaries where similarity drops significantly, confirmed by a long pause in the audio, are selected as chapter start points. For each chapter, the most prominent keyword phrases from the chapter content are used to auto-generate a title. You can edit any title before placing markers.
Click Place Markers to write chapter markers to your NLE timeline. Chapter marker names include the generated title so they're immediately readable in your NLE's marker panel. Click Export to generate the chapter list in your chosen format for copying into your video description, podcast CMS or Spotify.
AI-detected topic breaks with generated titles. Exportable to YouTube, Spotify and show notes. Zero manual timestamp copying.