VoiceLabs

Look up a VoiceLabs generation by id to poll its status. Returns status (generating | completed | failed) and, once completed, the audio URL; a failed one carries a short human-readable error and a failureKind. Read-only.

get_generation

Make sure a voice profile with the given name exists on the user's VoiceLabs account, creating it from one of VoiceLabs' built-in (preset) voices if it does not. Idempotent: calling it again with the same name returns the same profile with `created: false` and creates nothing. Use this when list_voice_profiles comes back empty, or when speak reports NOT_FOUND for a voice, then pass the returned id to speak. It cannot clone a real person's voice — that is clone_voice_profile, which needs the voice:clone scope and a recording.

ensure_voice_profile

Run a short TypeScript program that orchestrates VoiceLabs MCP tools (found via search_tools) in an isolated sandbox. See the server instructions for the full contract.

execute_typescript

Generate speech from text in one of the user's VoiceLabs voice profiles. Returns a generationId with status "generating"; poll get_generation with that id to retrieve the audio URL once it completes. Pass profileId or profileName to choose the voice.

speak

List the user's recent VoiceLabs captures (dictations, recordings, uploads) with their transcripts, most-recent first. Paginated via limit (1-200) and offset. Read-only.

list_captures

List the user's VoiceLabs voice profiles — both cloned voices and presets — with language, engine, and usage counts. Read-only. A brand-new account has none yet: an empty list comes back with an `emptyState` explaining how the user creates their first voice, which is what to relay instead of reporting that the tool failed.

list_voice_profiles

Find the VoiceLabs MCP tools this connection can reach, returned as TypeScript declarations to call from execute_typescript. See the server instructions for the full contract.

search_tools

Transcribe an audio file to text using VoiceLabs' Whisper. Give the audio exactly one way: audioUrl, a public https:// URL serving the file itself (wav, mp3, m4a, aac, ogg, flac, aiff or webm; up to 10 MiB), or audioBase64, the file's real bytes (≤10 MiB of base64 text). A payload that is not real audio — under 1 KiB, a bare header, or made-up placeholder bytes — is refused without being transcribed. If the user has an audio file you cannot read the bytes of, ask them for a public link, or to upload it on the Captures page of the VoiceLabs studio (https://app.voicelabs.now/studio). Returns the transcript and the created capture id.

transcribe