Hume MCP Server

@HumeAI

Hume MCP Server について

概要はまだありません

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "hume": {
      "command": "npx",
      "args": [
        "@humeai/mcp-server"
      ],
      "env": {
        "HUME_API_KEY": ""
      }
    }
  }
}

ツール

Synthesize (and play) speech from text

Replay previously generated audio

List available voices

Save a generated voice to your library

Remove a voice from your library

概要

What is Hume MCP Server?

The Hume MCP Server enables AI assistants to use Octave Text-to-Speech from Hume AI within chat interfaces. It integrates with MCP client applications like Claude Desktop, Cursor, or Windsurf, allowing developers to synthesize expressive speech that understands context, not just pronounce words.

How to use Hume MCP Server?

Add the provided JSON configuration to your MCP client's settings (e.g., claude_desktop_config.json), set your HUME_API_KEY environment variable, and install Node.js (a command-line audio player like ffplay is optional). The server exposes tools that can be called directly from the AI chat.

Key features of Hume MCP Server

Synthesize speech from text using Octave TTS.
Replay previously generated audio.
List, save, and delete voices in your library.
Configurable working directory for audio files.
Optional instant mode (10% extra cost) or embedded audio mode.

Use cases of Hume MCP Server

Narrate audiobooks or scripts with context-aware expressiveness.
Generate spoken dialogue for interactive storytelling or demos.
Manage and reuse custom voice profiles across projects.

FAQ from Hume MCP Server

What is Octave TTS and how is it different from traditional TTS?

Octave TTS is the first text-to-speech system built on LLM intelligence. It understands the meaning of words in context, delivering nuanced and expressive speech rather than flat pronunciation.

What dependencies or runtime does the server require?

You need a Hume AI account with an API key, Node.js installed, and optionally a command-line audio player such as ffplay (from FFmpeg) to play audio directly.

Where are generated audio files stored?

Audio files are saved to a working directory. By default this is the system temp directory under hume-tts, but you can override it with the --workdir command-line option or the WORKDIR environment variable.

Does instant mode affect cost?

Yes, enabling instant mode incurs an additional 10% cost on top of normal usage.

What authentication is needed to use the server?

The server requires the HUME_API_KEY environment variable set to your Hume AI API key. This is the only authentication mechanism described.

よくある質問

What is Octave TTS and how is it different from traditional TTS?

Octave TTS is the first text-to-speech system built on LLM intelligence. It understands the meaning of words in context, delivering nuanced and expressive speech rather than flat pronunciation.

What dependencies or runtime does the server require?

You need a Hume AI account with an API key, Node.js installed, and optionally a command-line audio player such as ffplay (from FFmpeg) to play audio directly.

Where are generated audio files stored?

Audio files are saved to a working directory. By default this is the system temp directory under `hume-tts`, but you can override it with the `--workdir` command-line option or the `WORKDIR` environment variable.