AllVoiceLab

AllVoiceLab MCP Server exposes a suite of AI-powered audio and video tooling via the MCP protocol. It enables voices generation, voice conversion, voice cloning, video translation and dubbing, subtitle removal, text translation, and subtitle extraction. Clients such as Claude Desktop, Cursor, Windsurf, or OpenAI Agents can connect to the AllVoiceLab MCP server to request text-to-speech, transform voices, clone a voice from samples, translate and dub videos, and generate subtitles. To start using it, ensure you have an API key from AllVoiceLab and configure the MCP client with the provided environment variables, including API key and domain. The server relies on uv (Python package manager) to run, so follow the installation steps to set up uvx and point it at the allvoicelab-mcp entry point.

Key capabilities include:

• text_to_speech: generate speech from text in 30+ languages with high realism.
• speech_to_speech: convert an existing voice to another voice while preserving content.
• clone_voice: create a new voice profile by cloning from an audio sample.
• remove_subtitle: remove hard subtitles from videos using OCR.
• video_translation_dubbing: translate and dub video speech into different languages while preserving tone.
• text_translation: translate text content into another language.
• subtitle_extraction: extract subtitles from videos via OCR.

To operate, provide your API key and domain in the env vars and connect via the MCP client using the AllVoiceLab server configuration. The tools will appear under the listed methods and can be invoked with natural prompts or structured requests, depending on your client’s capabilities.

Prerequisites:

• Python 3.x installed
• curl available to install uv script
• An AllVoiceLab API key and domain

Step-by-step installation:

1. Install uv (Python package manager):

   • Follow the official uv installation script: curl -LsSf https://astral.sh/uv/install.sh | sh

2. Install and configure the MCP server entry point:

   • The MCP server is exposed via the uvx command as shown in the configuration. There is no separate repository install beyond installing uv. The server binary is provided by uvx when you install uv.

3. Obtain API keys:

   • Get your API key from AllVoiceLab and record the API domain (global or regional as appropriate).

4. Configure environment variables:

   • In your MCP client configuration or environment, set: ALLVOICELAB_API_KEY=<insert-your-api-key-here> ALLVOICELAB_API_DOMAIN=<insert-api-domain-here> ALLVOICELAB_BASE_PATH=<optional path for outputs>

5. Run the MCP server via uvx:

   • Ensure the working directory is suitable and the environment variables are exported.
   • Example (shell): export ALLVOICELAB_API_KEY="<insert-your-api-key-here>" export ALLVOICELAB_API_DOMAIN="<insert-api-domain-here>" uvx allvoicelab-mcp

6. Connect with a MCP client:

   • Use the client’s configuration to point to the AllVoiceLab MCP server (e.g., Claude Desktop, Cursor) as shown in the example JSON config. Ensure Developer Mode is enabled if required by the client.

Notes:

• The Quickstart emphasizes region-specific API keys and domains; ensure you choose the correct domain corresponding to your region.
• If you encounter connectivity or tool availability errors, verify that the ALLVOICELAB_API_KEY and ALLVOICELAB_API_DOMAIN values are correct and that the network allows access to the API endpoints.

Tips and common issues:

• Environment variables are required for authentication and routing; do not omit ALLVOICELAB_API_KEY and ALLVOICELAB_API_DOMAIN.
• If you use Windows with Claude Desktop, enable Developer Mode to allow MCP servers.
• The ALLVOICELAB_BASE_PATH is optional but recommended for file outputs, especially large audio/video files.
• Permissions: ensure the user running uvx has permission to read/write to the base path.
• If you switch regions, generate a new API key for the corresponding region and update ALLVOICELAB_API_DOMAIN accordingly.
• MCP client tools expose methods like text_to_speech, video_translation_dubbing, and subtitle_extraction; consult the AllVoiceLab documentation for exact prompts and parameter usage.
• Logs: monitor MCP logs at the standard location for your OS; check for network-related or API quota issues if tasks hang or fail.