Speak for Openclaw

A local, high-performance text-to-speech engine powered by Kokoro for AI agents.

babysor
v1.0.0
Mar 2, 2026
0
845
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install sayit

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install sayit using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Speak?

The Speak skill provides AI agents with the ability to generate natural-sounding speech locally using the Kokoro TTS engine. Unlike cloud-based solutions, this tool runs entirely on your hardware, ensuring privacy and low latency for all voice synthesis tasks. As a featured utility within the Openclaw Skills collection, it enables developers to build sophisticated audio-driven interfaces without relying on external API subscriptions.

By leveraging ONNX model acceleration, the skill produces high-fidelity audio across multiple languages and voices. It is designed to be lightweight yet powerful, fitting seamlessly into automated workflows where text needs to be transformed into audio files or streamed in real-time.

Speak Use Cases

  • Creating audio versions of technical documentation or articles for better accessibility.
  • Building responsive AI assistants that communicate via localized speech rather than just text.
  • Converting long-form content like EPUB or PDF books into structured audiobooks with chapter divisions.
  • Developing privacy-centric applications that require voice output without sending data to third-party servers.

How Speak Works

  1. The AI agent identifies text content that needs to be converted into speech.
  2. The Speak skill invokes the local Kokoro engine, passing parameters such as voice selection, language, and playback speed.
  3. The engine processes the input using localized ONNX models and voice binary files.
  4. Audio is generated either as a file (WAV/MP3) or as a live stream for immediate playback.

Speak Setup

To integrate this capability into your Openclaw Skills environment, install the tool via uv and download the necessary model weights:

uv tool install kokoro-tts
wget https://github.com/nazdridoy/kokoro-tts/releases/download/v1.0.0/kokoro-v1.0.onnx
wget https://github.com/nazdridoy/kokoro-tts/releases/download/v1.0.0/voices-v1.0.bin

Speak Data Schema & Taxonomy

The skill utilizes a structured approach to managing voice assets and outputs:

Feature Details
Core Models kokoro-v1.0.onnx and voices-v1.0.bin
Supported Input String literals, .txt, .epub, .pdf
Output Formats .wav, .mp3, or stdout streaming
Voice Taxonomy Categorized by Region (US, GB, CN, JP, FR, IT) and Gender (Male/Female)

Speak Advanced Features

  • Voice Blending: Create unique vocal personas by mixing existing voices with weighted ratios.
  • Multi-Language Support: Native synthesis for English, Mandarin Chinese, Japanese, French, and Italian.
  • Speed Control: Dynamically adjust the pacing of speech for different contexts.
  • Batch Processing: Automatically split large documents into individual audio chapters for easier management.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*