Text to Speech

Text to Speech reads a document aloud from start to finish and saves the result as one MP3 in your library. It is not a summary and not a podcast -- no hosts, no discussion, no rewriting. You get the document itself, narrated, so you can listen to it on a commute or while you work.

Where to Start

There are two ways in:

• From Home -- pick Read Aloud from the tool row above the prompt box.
• From a source -- open a file or PDF and choose Text to Speech from its AI creation actions, under Listen. The source you were looking at comes in already selected.

Both open the same Text to Speech window.

Choosing What to Read

Pick sources from your library, or upload a PDF, Word document, PowerPoint, or text file. You can also paste a PDF or website URL, or pull a file straight from Google Drive -- see Connected Apps.

A few things worth knowing before you pick:

• Up to five sources can go into one narration. When there is more than one, each document after the first is announced by title so they don't run together mid-sentence.
• PDFs can be narrowed to a page range. Use the page selector on the customize step to read a chapter instead of the whole book -- this also lowers the cost.
• Formatting characters are stripped before narration, so the voice reads prose rather than saying "hash hash Introduction" at every heading.

Voice and Speed

On the customize step you choose how it sounds.

Voice

Click the voice button to open the narrator list. It is searchable, and every voice has a play button so you can hear a short demo before committing. If you'd rather not choose, leave it on Default voice -- a clear, neutral reader picked for long documents.

Speed

Pick 0.75×, 1×, 1.25×, or 1.5×. The speed is baked into the audio itself rather than applied at playback, so a faster narration is genuinely a shorter file.

Language

The Language row uses your shared language preference, the same one every other create window reads.

What It Costs

Narration is priced by length: 1 AI credit per 10,000 words, up to 30,000 words per narration.

You see the exact price before anything is charged. Once your sources are selected, the Cost panel measures them and shows:

• the credit cost for this narration,
• roughly how many minutes of audio you will get at your chosen speed,
• the measured word count.

The create button carries the same number, so it reads Create Narration · 2 AI credits rather than leaving you to guess. Edit the page range and the quote re-measures.

If your selection is over the 30,000-word limit, the window tells you the word count and blocks the run. Remove a source or narrow the page range and try again. If you don't have enough credits for the quoted cost, it says how many you have and offers the upgrade path -- see Plans and Limits.

While It Runs

Narration runs in the background. Long documents take a few minutes, and you don't need to wait on the screen -- close the window, keep working, and Scholarly opens the finished audio when it is ready. You can follow progress from the background tasks strip on Home. See Background Tasks and Notifications.

The Finished Audio

The MP3 lands in your library and opens on its own audio page, showing the duration, file size, and date under the title. Click the title to rename it.

Below the player are two tabs:

• Transcript -- the full narration text, with a copy button. Click any line to jump the player to that moment.
• AI & Sources -- what this narration was made from, and what else has been created from that same source.

Audio pages don't offer create actions of their own -- a narration is an output. To build flashcards, a study guide, or anything else, go back to the original document and create from there.

Tips and Limits

• Narration reads documents, not typed prompts. There is no blank-page start -- you always pick a source.
• Supported uploads are PDFs, Word documents, PowerPoints, and text files. Other formats are rejected in the picker.
• A long book is usually better narrated a chapter at a time: cheaper, faster, and easier to navigate later.
• If a narration fails, the credits it reserved are refunded.

Related

• Uploading Content -- getting your documents into Scholarly.
• AI Podcasts -- a multi-speaker discussion about your sources, rather than a straight read.
• Library Files -- how files and generated outputs are stored and organized.
• Plans and Limits -- what AI creation credits cover on each plan.