Preparing for narration
What will sound wrong when your book is read aloud: web addresses, tables, abbreviations, Roman numerals, acronyms. Opens your Word and EPUB files.
This application is served by KNPLIV and runs sandboxed: it cannot read your sign-in details, store anything on your device, or send in the background what you type into it. Everything is gone when you leave the page. Like everything published here, it was read over before going online.
Amazon narrates an ebook for free with its virtual voice. What an author cannot do alone is spot, across three hundred pages, what will sound wrong.
This tool goes through your text the way a synthetic voice would and points out nine things: web and email addresses, which a voice spells out character by character; tables, which read as strings of unconnected words; footnote markers, surfacing mid-sentence; Roman numerals, where “Chapter XIV” becomes “ksiv”; abbreviations; unfamiliar acronyms; symbols and units; scene breaks, heard as “asterisk asterisk asterisk”; and sentences so long the listener loses the thread, unable to glance back.
Every doubtful passage carries its line number and a button to hear it, using the voices installed on your device, at the rate and pitch you set.
It corrects nothing on your behalf: an automatic fix would get the right cases wrong, and you would no longer know what had moved. It shows, you decide.
Nothing leaves your browser: your text is sent nowhere, and no voice is fetched from the internet.