Command: /create-voice
Relevant source files
The following files were used as context for generating this wiki page:The /create-voice command is a feature of the Voice ElevenLabs plugin that allows users to generate unique, custom AI voices directly from a text description. It leverages the ElevenLabs Voice Design v3 feature, eliminating the need for audio samples. This enables users to create personalized voices for use with other plugin commands, such as /speak or within the voice-assistant agent.
This command is part of a suite of tools that add natural voice capabilities to the Claude Code environment, including text-to-speech, dialogue generation, and transcription. Once a voice is created, it is saved to the user's ElevenLabs account and can be managed using the /voices and /voice-config commands.
Sources: README.md:6, 32, commands/create-voice.md:2, 6-10, agents/voice-assistant.md
Command Flow
The /create-voice command follows a straightforward process that starts with a user's description and ends with a new voice ID being available for use. The plugin orchestrates the interaction with the ElevenLabs API.
Sources: commands/create-voice.md:5-10, 37-42
The diagram below illustrates the end-to-end process of creating a new voice.
Sources: commands/create-voice.md:6-10, 37-42
Usage
To use the command, a user provides a natural language description of the desired voice.
Sources: commands/create-voice.md:5
Syntax
/create-voice "[description]"
Examples
Simple description:
/create-voice "warm female narrator, mid-30s, slight British accent, professional"
Sources: README.md:33, commands/create-voice.md:22-23
More detailed description:
/create-voice "middle-aged New Yorker with rising intonation and a half-smile in his voice, like a friendly cab driver telling stories"
Sources: commands/create-voice.md:31-32
With audio quality specifications:
/create-voice "energetic young voice, gender-neutral, upbeat and enthusiastic, podcast host style, perfect audio quality"
Sources: commands/create-voice.md:27-28, 35
Description Guidelines
The quality of the generated voice depends heavily on the specificity of the description. The command documentation recommends including details about several voice characteristics.
Sources: commands/create-voice.md:13-14
| Category | Description | Examples |
|---|---|---|
| Age | The perceived age of the speaker. | young, middle-aged, elderly |
| Gender | The gender identity of the voice. | male, female, neutral |
| Accent | The regional or national accent. | British, Southern US, Australian, French |
| Tone | The emotional or stylistic tone. | warm, authoritative, playful, serious |
| Quality | The physical quality of the voice. | clear, raspy, smooth, deep |
Sources: commands/create-voice.md:15-19
Audio Quality Control
Users can also specify the desired audio recording quality in the description.
"perfect audio quality"for a clean, studio-like sound."slight room reverb"for a more natural, ambient feel."radio effect"for a vintage sound profile.
Sources: commands/create-voice.md:34-36
Implementation and Post-Creation
The command is defined to use Bash and Read as its allowed tools. The core logic involves calling the ElevenLabs Voice Design v3 API, processing the results, and saving the selected voice.
Sources: commands/create-voice.md:3, 37-42
Key Steps
- Receive the voice
descriptionand an optionalnamefrom the user's arguments. - Use the ElevenLabs API to generate three voice candidates based on the description.
- Present the options to the user, including audio samples for preview.
- Save the user's selected voice to their ElevenLabs account.
- Return the new
voice_idto the user. This ID can then be used in other commands or configurations.
Sources: commands/create-voice.md:37-43
Managing the New Voice
Once a voice is created, its ID is returned. This ID can be used to:
- List all available voices, including custom ones, via the
/voicescommand. - Set the new voice as the default for the session or globally by using the
/voice-config set-voice [voice_id]command or by setting theELEVENLABS_VOICE_IDenvironment variable.
Sources: commands/voices.md:17-19, commands/voice-config.md:13, 19
The new voice can then be utilized by the voice-assistant agent or the voice-responses skill to provide spoken output.
Sources: agents/voice-assistant.md:19, skills/voice-responses.md:14