Voices

A video's voice comes from its preset. To change it per video, send voice:

"voice": { "id": "elevenlabs:21m00Tcm4TlvDq8ikWAM", "speed": "normal" }

Voice IDs

GET /v1/voices (MCP: list_voices) lists every voice you can use: the voice library, PRO voices, your own cloned voices, your favourites and the voices your team shares with you. Each has:

• id — "<service>:<voice>"; send it as voice.id exactly as listed.
• name, gender, description, preview_url.
• languages — the languages it speaks; an empty list means a multilingual voice.
• source and shared_by — where it comes from.
• available — false when it needs your own provider key (Settings → API Keys) that this account doesn't have.

Filter the list:

curl "https://api.vidnext.ai/v1/voices?language=uk&q=olena" -H "Authorization: Bearer $VIDNEXT_API_KEY"

provider= narrows it to one voice service, limit and offset page through it.

Language and speed

• A voice must be able to read the video's language. A voice made for another language is refused with voice_language_mismatch; send "allow_language_mismatch": true to use it anyway (it reads with its own accent).
• Some voice services read only some languages; a voice whose service can't read the video's language is refused with tts_language_unsupported.
• speed (slowest, slow, normal, fast, fastest) works on the voices that support it — GET /v1/options → narration_speeds says which.

Your own narration

Send your text as script and the voice reads it word for word — see Create a video.