Prepare a document
Start with a saved document in ScreenApp. If your material is a recording, first generate and review a document from it. If you start with pasted text or an imported document, check that the saved document contains the wording you want spoken.
Remove editor comments and instructions that should not be read aloud. Spell out an ambiguous abbreviation where pronunciation matters.
Request the audio
- Open the document or attach it in Ask AI.
- Ask: “Read this document aloud in English, keeping its wording and section order.”
- Replace English with the language you want.
- Wait for the returned audio file to finish processing.
- Play the result before downloading or sharing it.
The document read-aloud tool creates a spoken audio file. If you want a shorter adaptation, first ask for and approve a shorter document, then have that document read aloud. This makes it easier to check what was omitted.
Review the narration
Listen to names, numbers and specialist terminology. Compare section transitions with the written source. Tables, links and citations can need rewriting for listening because their visual structure does not translate directly into speech.
Synthetic speech can mispronounce a word even when the written text is correct. Revise the source wording where appropriate and review the next output instead of assuming a retry fixed it.
Download or share
Open the returned audio in your library and use its download action. The source document remains useful as a written reference. Check download access if the app blocks the download.
For an existing recording that needs speech in another language, use dubbing. For the exact document workflow and failure checks, see read a document aloud.