Model Context Protocol moderate risk

Voice

VoiceMode enables natural voice conversations with Claude Code, offering hands-free interaction through speech-to-text and text-to-speech capabilities, with local and remote options.

Connections & Capabilities

Connects To

GitHub

Capabilities

readwriteexec

Quickstart

Install

brew install ffmpeg node portaudio

Config

{
  "mcpServers": {
    "voice-mcp": {
      "command": "uvx",
      "args": ["voice-mcp"],
      "env": {
        "OPENAI_API_KEY": "your-openai-key"
      }
    }
  }
}

Exposed MCP Tools (6)

moderate

/voicemode:install

Installs dependencies and local voice services required for VoiceMode.

Installs system packages and configures environment.

moderate

/voicemode:converse

Starts a voice conversation with Claude Code.

Allows interaction with the agent, potentially leading to unintended actions based on voice commands.

high

voicemode agent start

Starts the operator agent in a tmux session, enabling remote access.

Exposes the system to external connections and potential unauthorized access.

moderate

voicemode agent stop

Stops the operator agent.

Kills the tmux window, interrupting any ongoing processes.

moderate

voicemode agent send

Sends a message to the operator agent, potentially triggering actions.

Can initiate actions based on the content of the message.

high

voicemode config edit

Opens the configuration file for VoiceMode, allowing modification of settings.

Incorrect configuration can lead to security vulnerabilities or unintended behavior.

Safety Assessment

VoiceMode offers a balance between convenience and security. Using local voice services and carefully configuring permissions can mitigate risks, while enabling remote agent functionality and relying on cloud services introduces potential vulnerabilities.

Optional local voice services for enhanced privacy.
Remote agent functionality can be disabled.
Permissions setup allows for controlled access to features.
Configuration options for customization and security.

Requires microphone access, which could be a privacy concern.
Remote agent functionality exposes the system to external connections.
Use of cloud services (OpenAI) depends on their security practices.
Improper configuration could lead to unintended access or exposure.