What is Voice Match?

Voice Match lets you save a Writing style for prose or a Speaker style for a particular person's dialogue or first-person speech. Create each style by writing your own guide or asking AI to analyze text. Each saved style contains one guide, which you can select directly in AI features.

Speaker styles are in beta.

Note

Writing your own guide does not call AI. Analyzing a sample uses your selected AI model and may consume managed AI usage.

Creating a Voice

To create a profile:

Open Workspace Settings and find Writing styles.
Click New voice and choose Writing style or Speaker style.
Name the style, then choose Analyze with AI to provide text or a URL, or Write my own to save your guide directly. For a speaker style, you can optionally enter the speaker to find in the text.
Create the Voice.

The writing-style analysis examines tone, cadence, sentence structure, word choice, pacing, and descriptive style. The speaker-style analysis focuses on dialogue: diction, rhythm, slang, connotations, quirks, and ways of addressing others. Leave Speaker to analyze blank to analyze the dialogue provided.

You can edit a guide you wrote. For an analyzed style, Analyze again lets you review a new analysis before replacing the current one.

Tip

Use samples that clearly represent the style or speaker you want to capture. An analysis of a short or mixed-speaker sample may be incomplete.