Descript's AI stock speakers allow you to easily create lifelike voiceover audio by simply typing text into your project. You can even add multiple stock speakers to your project to create full dialogues using AI speech.
Adding a stock speaker to a project
To create text-to-speech using one of our stock voices, you will first need to assign a stock speaker to the script. You can do so by first creating a new project or open an existing one.
Next, place your cursor on a blank line in your script wherever you want to use a stock speaker. If this is a brand new project, press Enter to continue with an empty composition.
If you already added and transcribed video or audio in you script, just press Enter (Windows) or Return (macOS) to create a new line.
Assign a speaker label by typing @ and selecting browse stock AI speakers at the bottom of the speaker list.
Alternatively, if you have already written some of your script using Write mode, you can click the Add speaker button above the section of blue placeholder text.
Choose any stock speaker from the list to add it to your project; you also can preview each speaker by clicking the play to the right of a speaker's name.
We updated our AI Speaker model in October 2023. Since then, we've improved and retrained our stock AI speakers using our new AI model. Any stock speaker with 'Legacy' next to its name uses the old AI model.
If you would still like to access or use these legacy stock speakers, you can show them by clicking the filter icon and changing the filter from Hide to Show.
You will not have legacy speaker options if you created your Descript account after the switch to the new model.
Stock speakers Usage
There are important considerations to be aware of when using stock speakers in your projects:
- For now, stock speakers are only available in English.
- Stock speakers are only for generating text-to-speech in your projects. They cannot be used for overdubbing audio.
- If you want to change the stock speaker for a section, just click on the stock speaker name above the text, and choose another stock speaker. The new speaker's audio with begin generating once you swap in the new speaker.