Did this help?
Narrator Friend is free and stays free. If it saved you time, a small Venmo tip keeps the lights on.
Tip via Venmo · @Amy-barthelemy12A few notes before you trust the numbers
- Counts are estimates, not gospel. Dialogue / narration / tag classification is based on text patterns and won't be 100% accurate, especially in poetry, scripts, or unusually formatted prose. Treat the recording-time number as a planning estimate, not a quote-locking guarantee.
- Chapter breaks need to be recognizable. The parser uses your file's bookmarks (PDF) or Heading 1 styles (Word), and falls back to heading patterns like
Chapter 1/CHAPTER ONE/Prologue. If your manuscript runs chapters together with no headings, the whole book may be reported as one chapter. - Scanned PDFs need OCR first. If your PDF is a picture of pages rather than searchable text, there's nothing for the parser to read. OCR converts the image back into text. We don't run OCR here, but you can do it locally before uploading: Adobe Acrobat (Tools › Scan & OCR), OCRmyPDF (free, open source, runs on your own machine), or Sejda's free online OCR. None of these are AI / LLM tools.
- POV detection is conservative. It only fires when the chapter title or first lines look unambiguously like a name (e.g.
Chapter 5: Sarah, or aSARAHsubheader).