Voices and narration

Choose a Poko voice or connect a supported voice provider.

Last updated September 2026

Choose a voice

Set a narration voice from the Voice tool in Studio or from Settings → Voices. The selection and active provider save automatically, and narration uses the saved voice for that provider.

Bring your own provider

Cartesia, ElevenLabs, and VocalLab catalogs become available after you save the relevant API key in Settings → Voices. Poko does not load a provider’s catalog until its key is configured.

Voice availability

A selected voice must include a valid language before narration can run. Voice and model settings are locked while an agent session is running, so make selections before starting a generation.

Create a voice clone

Open the Poko Clone tab, give the voice a name and language, then record a sample or upload audio. Confirm that the voice is your own or that you have explicit written permission before creating the clone.