M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
V

VoiceMode

Community
Natural voice conversations with Claude Code and other MCP-capable AI assistants.
How to install ↓ Compare GitHub source repository ↗
Category
Dev Tools #176 of 438
Stars
★ 1.4k Very popular
Transport
stdio (local process)
Runtime
Python 3.10-3.14 · Prebuilt binary
Credentials
Optional API key
License
MIT
Last commit
Tools
2
56FMRS · C

VoiceMode is a practical MCP server that adds natural voice interaction to Claude Code. It supports both cloud and local speech processing, with sensible defaults like silence detection. Setup is straightforward via the provided installer, and it is well-suited for hands-free coding workflows. Privacy-conscious users should note that cloud STT/TTS sends audio to third-party servers, while local models require substantial system resources. Overall, it is a mature and actively maintained open-source project.

Strongest · Setup experience 13/20 Weakest · Reliability 8/20

Reliability
8/20
Security and permissions
10/20
Maintenance
12/20
Documentation
13/20
Setup experience
13/20
Why each score
Reliability 8/20
Only the README and manifest were reviewed; no source code, tests, or CI results are available. The manifest is well-formed and consistent with the README (voice-mode package, stdio transport, launcher), but no tool schemas or execution evidence could be verified. No tests or CI workflows are present, so uncertainty is high; score is 8.
Security and permissions 10/20
No red-line issues found: no hardcoded secrets, no malware, no covert exfiltration instructions; OPENAI_API_KEY is optional and marked secret. However, without source code we cannot verify least-privilege behavior, data boundaries, or confirmation flows. The README suggests optional permission overrides, which is user-controlled; deductions for unverifiable data-flow disclosure and permission scoping give a score of 10.
Maintenance 12/20
Repository is not archived, has an MIT license, a recent version 8.12.0, and an issue tracker with 24 open issues, suggesting active maintenance. However, no commit history, release dates, dependency-update evidence, or security-response channel is present in the supplied material; stars are not counted. Missing verifiable signals yields 12.
Documentation 13/20
The README covers quick start, installation alternatives, compatibility, configuration, troubleshooting, and links to further guides; the manifest exposes environment variables with descriptions. But it lacks explicit tool parameter documentation, limits, cost information, and linked documentation cannot be verified. Score is 13.
Setup experience 13/20
The README provides concrete setup paths (Claude Code plugin and uvx installer), including client configuration examples and platform dependencies. No CI/test evidence exists in the supplied data, so the ceiling is 15; the process is also manual and has multiple system dependencies, making it fragile in unknown environments. Score is 13.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessRuns commands or codeUses the network

Best for

  • Claude Code users who want to speak to their coding assistant
  • Developers who need hands-free interaction during tasks
  • Users who prefer local, private speech processing

Not for

  • Users without a microphone or speakers
  • Text-only environments where audio is not available

Required permissions

  • Microphone access for capturing voice input
  • Speakers or audio output devices for playing responses
  • Network access when using cloud STT/TTS services (e.g., OpenAI)
  • Claude Code MCP permissions for the converse and service tools

Risks and side effects

  • When cloud speech services are enabled, audio is transmitted to third-party servers (e.g., OpenAI)
  • Local speech models (Whisper, Kokoro) can be resource-intensive
  • Microphone access means the MCP process captures ambient audio

Setup

Before you start

Runtime：Python 3.10-3.14 · Prebuilt binary

OPENAI_API_KEY optionalsecret Optional; cloud STT/TTS fallback when local services aren't installed. Get it from the OpenAI platform.
Other optional settings (6)
VOICEMODE_DEBUG optional Set to true to enable debug mode with detailed logging.
VOICEMODE_SKIP_TTS optional Set to true to skip TTS and show text only for faster responses.
VOICEMODE_PREFER_LOCAL optional Set to true (default) to prefer local services over cloud when available.
VOICEMODE_AUDIO_FORMAT optional Audio format: pcm, mp3, wav, flac, aac, opus; default is pcm.
VOICEMODE_WHISPER_MODEL optional Local Whisper model: tiny, base, small, medium, large; default is base.
VOICEMODE_DISABLE_SILENCE_DETECTION optional Set to true to disable silence detection for continuous recording.
  1. Install the UV package manager if needed: curl -LsSf https://astral.sh/uv/install.sh | sh
  2. Run the installer: uvx voice-mode-install
  3. Add to Claude Code: claude mcp add --scope user voicemode -- uvx --refresh --from voice-mode voicemode-mcp-launcher
  4. Optionally set OPENAI_API_KEY for cloud services.
  5. Start a conversation: claude converse
.mcp.json 
{
  "mcpServers": {
    "voicemode": {
      "command": "uvx",
      "args": ["--refresh", "--from", "voice-mode", "voicemode-mcp-launcher"]
    }
  }
}

Shown for Claude Code. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json
{
  "servers": {
    "voicemode": {
      "command": "uvx",
      "args": [
        "--refresh",
        "--from",
        "voice-mode",
        "voicemode-mcp-launcher"
      ]
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal
claude mcp add voicemode -- uvx --refresh --from voice-mode voicemode-mcp-launcher

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

In Claude Code, run /mcp and confirm the voicemode server is connected with the converse and service tools listed; then run /voicemode:converse to start a spoken conversation as an end-to-end check.

Troubleshooting

  1. No microphone access: Check terminal/app permissions. On WSL2, install pulseaudio packages.
  2. UV not found: Run curl -LsSf https://astral.sh/uv/install.sh | sh.
  3. OpenAI API error: Verify that OPENAI_API_KEY is set correctly.
  4. No audio output: Check system audio settings and available devices.

Things to try

Once connected, you can ask your AI assistant things like:

  • Let's talk by voice — use my microphone to discuss today's to-do list
  • Check whether my local Whisper speech service is running, and start it if not
  • Switch to Kokoro local text-to-speech, then read a paragraph aloud to me
  • Skip TTS and just show text so responses come back faster

Tools 2

converse read-only
Start a natural voice conversation with the AI assistant.
service writes
Manage the VoiceMode voice service.

Use cases

Walking to a meeting while keeping up with work
Cooking while debugging code
Giving your eyes a break after long screen time
Holding a coffee or a dog while coding
Hands-free operation when typing isn't possible

Supported clients

Claude Code

Listed from the project's documentation, not tested by this site.

Overview

VoiceMode is an MCP server that brings natural voice conversations to Claude Code and other MCP-capable agents. It provides speech-to-text and text-to-speech via the MCP protocol, with optional local models (Whisper, Kokoro) or cloud services (OpenAI). It features low latency, smart silence detection, and works offline when local services are installed.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 63.3k · Tools 2 Compare with this →
Compare with other MCP servers

Source revision e03b5d91a984 Data synced 2026-10-19 Read the FMRS scoring method