Turn a manuscript into manageable voice jobs
The splitter recognizes headings such as Chapter, Part, Section, and Markdown headings. It shows each title and character count, generates a selected chapter of up to 1,000 characters as a local WAV, and exports a JSON manifest containing the chapter text. Longer chapters should be divided into scenes at paragraph boundaries.
Why chapter boundaries matter
Short, named jobs are easier to regenerate when a pronunciation or sentence changes. Keep front matter, chapter titles, body text, notes, and credits separate. Decide whether titles should be spoken, and use stable filenames that sort in listening order. A chapter can still be too large for one browser generation, so divide long chapters into scenes at paragraph boundaries.
Audiobook quality and rights
Synthetic narration does not replace editing. Listen for missing paragraphs, duplicated text, abbreviations, dialogue attribution, and abrupt joins. Keep the manuscript outside the browser as the source of truth. Before publishing, verify your rights to the text and the selected voice or model, and do not imply that a generated character is a real person.
Questions people ask
Does this tool upload my content?
No. Tool input is processed in the current browser. Required code or model files can be downloaded from the providers disclosed in the privacy policy.
Do I need an account?
No account or subscription is required. Browser and device limits still apply.
Can I use the result commercially?
This website does not grant rights to text, recordings, device voices, models, or third-party assets. Check every applicable license for the intended use.