What is VLab

VLab (Voice Lab) is a text-to-speech bridge for AI coding agents. Instead of silently finishing a long task and hoping you notice, your agent can speak — announcing completions, surfacing errors, asking for your attention — without interrupting your flow.

It exposes a single MCP tool (speak) that any MCP-compatible agent can call. The agent passes text; VLab speaks it aloud through xAI Grok or ElevenLabs voices on your Mac (with a free offline say fallback). You set the persona, mood, and snark level. You can mute it or stop mid-sentence when you're on a call.

Requirements: macOS Apple Silicon · an xAI (Grok) and/or ElevenLabs API key

Install

# 1. Install vlab
curl -fsSL https://vlab.daystra.com/install.sh | sh

# 2. Store a TTS API key in your macOS Keychain (set at least one).
#    vlab reads these by service name from the Keychain — no env vars,
#    no secrets in dotfiles.

#    xAI Grok  — default voices: rex, ara, leo, eve, sal
security add-generic-password -s GROK_API_KEY -a "$USER" -w "your-grok-key"

#    ElevenLabs — voices: veda, alfred, + any you adopt
security add-generic-password -s ELEVENLABS_API_KEY -a "$USER" -w "your-elevenlabs-key"

# 3. Verify everything's wired (keys, audio, config)
vlab doctor

# 4. Wire vlab into Claude Code (registers the MCP server + installs the skill)
vlab install

# 5. Speak from your terminal
vlab speak "Build complete. All tests passed."

Tip: vlab doctor tells you which keys it found. Both services are optional individually — set whichever voice engine you want. (Env vars $GROK_API_KEY / $ELEVENLABS_API_KEY also work if you prefer.)

Features

MCP tool: speak — agents call it with text; VLab converts via xAI Grok or ElevenLabs and plays audio locally. Works from Claude Code, Cursor, Windsurf, or any MCP client.
Personas — pick a voice character: Grok's rex, ara, leo, eve, sal or ElevenLabs' veda, alfred. Set globally or per-call.
Mood & snark dials — tune the delivery. Set the mood (neutral, upbeat, urgent) and snark level (0–10) so the voice matches the moment.
Mute / stop — vlab mute silences output; vlab stop cuts the current utterance mid-sentence. Both take effect immediately.
Self-update & one-command setup — vlab update upgrades to the latest release in place; vlab install wires the MCP server + skill into Claude Code for you.
Doctor check — vlab doctor verifies your Grok / ElevenLabs keys, audio device, and config in one command.

Wiring into Claude Code

One command registers the global MCP server and installs the agent skill into ~/.claude:

vlab install

Then restart Claude Code. Prefer to wire it by hand (or use another MCP client)? Add this to your project's .mcp.json:

{
  "mcpServers": {
    "vlab": {
      "command": "vlab",
      "args": ["mcp-server"]
    }
  }
}

Once wired, your agent can call vlab__speak any time it wants to surface something audibly — task done, build failed, waiting for your input.

CLI Reference

vlab speak "<text>" — speak text immediately
vlab say "<text>" — alias for speak
vlab mute / vlab stop — silence output / cut the current utterance
vlab install — wire the MCP server + skill into Claude Code
vlab update — upgrade VLab to the latest release
vlab voices [--search <q>] — list voices or search the library
vlab usage — this month's TTS character spend
vlab doctor — check keys, audio device, and config
vlab mcp-server — start the MCP server (used by .mcp.json)
vlab --version — print installed version

macOS (Apple Silicon) · Apache-2.0 / MIT · Requires an xAI Grok or ElevenLabs API key · llms.txt

A Cybercussion Interactive, LLC project