Speakturbo Tts

This is the TTS skill you'd actually use during development because it's stupidly fast. 90ms to first sound versus the usual 4-8 second wait means you can get audio feedback without breaking your flow. The tradeoff is simple: you get 8 built-in voices instead of custom cloning, and no emotion tags. First run takes a few seconds while the daemon loads a 100MB model into memory, then it stays warm for an hour. The Rust CLI streams from a Python daemon running pocket-tts, and it'll auto-manage the daemon lifecycle. If you need Morgan Freeman or laugh tags, use the speak skill instead. If you just need Claude to talk back without the wait, this is it.

skills.sh installs: 1K
GitHub stars: 20
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add emzod/speak-turbo --skill speakturbo-tts --agent claude-code

Installs into .claude/skills of the current project.

Files

speakturbo - Talk to your Claude!

Give your agent the ability to speak to you real-time. Ultra-fast text-to-speech with ~90ms latency and 8 built-in voices.

Quick Start

# Play immediately - you should hear "Hello world" through your speakers
speakturbo "Hello world"
# Output: ⚡ 92ms → ▶ 93ms → ✓ 1245ms

# Verify it's working by saving to file
speakturbo "Hello world" -o test.wav
ls -lh test.wav  # Should show ~50-100KB file

Output explained: ⚡ = first audio received, ▶ = playback started, ✓ = done

First Run

The first execution takes 2-5 seconds while the daemon starts and loads the model into memory. Subsequent calls are ~90ms to first sound.

# First run (slow - daemon starting)
speakturbo "Starting up"  # ~2-5 seconds

# Second run (fast - daemon already running)
speakturbo "Now I'm fast"  # ~90ms

Usage

# Basic - plays immediately (default voice: alba)
speakturbo "Hello world"

# Save to file (no audio playback)
speakturbo "Hello" -o output.wav

# Save to specific file
speakturbo "Goodbye" -o goodbye.wav

# Quiet mode (suppress status messages, still plays audio)
speakturbo "Hello" -q

SKILL.md on GitHub