Synthetic voices

Generate narration from your own voice library.

Prism generates narration in the output's presenter voice for any block without a recording. Most scripts are narrated this way, and it needs no setup.

The presenter comes from the brief

You choose a presenter at the Assignment Desk, in the PRESENTER slot, before anything is created.

It is required. Until a presenter, a script style and a video style are all set, both create buttons stay disabled.

Presenter cards show a name and a mode. The two modes are Personality in Script and Voice Only.

Your choice follows the output into the Writers Room, where each block shows the presenter and mode above its text.

Your first video walks through briefing an output from an empty project.

Filling the gaps when you commit

If you have uploaded recordings and some blocks are not covered, Missing voiceover coverage opens when you commit.

It lists every uncovered block with a dropdown, set to Generate with AI by default.

• Leave a block on Generate with AI to have Prism narrate it
• Switch it to Upload file… to supply a recording instead
• Switch it to Record in browser… to record one there and then
• Switch it to Skip block to leave it out

Then choose Commit with AI fill to go ahead, or Go back to return to the script.

An uploaded recording always wins. Prism does not generate audio for a block that already has one.

Voices are your workspace's own

Prism ships no fixed list of voices, and no fixed list of languages.

Both come from your account. The voices in the pickers are the ones connected to it, and you can connect your own ElevenLabs voice library under CONNECTORS.

The output's language is a brief setting, shown as a badge at the top of the Writers Room.

Building a custom presenter

Custom presenters are made in Manage Voices, reached from the voice connector under CONNECTORS in the account menu.

1. Choose Add voice
2. Pick a voice from your connected voice library
3. Give the presenter a name that says what it is for
4. Adjust Voice settings if the voice needs it
5. Choose Create presenter

Voice settings offers sliders for Stability, Similarity, Style and Speed, a Speaker boost toggle, and a Model dropdown.

Leave these alone unless a read sounds wrong. The saved settings behind a voice are usually the right starting point.

Testing before you use it

Test voice speaks a line of your choosing so you can hear the result.

Type the text, choose Test, then play the sample back.

Testing is unavailable until the presenter has been saved once, and the hint on the control says so.

Saved presenters are listed in Manage Voices with a sample you can play, an edit control and a delete control.

Deleting is permanent, and you can only delete presenters you own. Shared ones stay put.

FAQ