A high-performance, 100% offline text-to-speech skill for generating natural voice replies without API keys or internet dependencies.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install voice-reply
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install voice-reply 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 Voice Reply?
Voice Reply is a privacy-first text-to-speech (TTS) solution designed to give your AI agent a clear, natural voice without relying on cloud-based providers. By utilizing the sherpa-onnx runtime and Piper voice models, this skill provides high-speed audio generation entirely on your local hardware. It is a vital component for developers building Openclaw Skills that require secure, low-latency vocal responses.
This skill is specifically optimized for Linux environments and includes seamless integration with messaging platforms like Telegram. It transforms plain text into high-quality OGG Opus files, automatically tagging them for display as interactive voice bubbles. Whether you are building a private assistant or a localized automation tool, Voice Reply ensures your agent can communicate audibly while keeping all data on your own machine.
Voice Reply Use Cases
- Sending hands-free voice notes via Telegram for mobile accessibility.
- Creating localized voice notifications for home automation or server alerts.
- Operating AI agents in air-gapped or privacy-sensitive environments where cloud TTS is prohibited.
- Providing multi-language support (German and English) for international user bases.
- Reducing operational costs by eliminating per-character API fees from external speech providers.
How Voice Reply Works
- The skill receives a text input from the AI agent, along with an optional language identifier.
- It performs an automatic language detection check if the language is not explicitly specified (detecting German via umlauts or English as a default).
- The text is passed to the sherpa-onnx offline engine which loads the corresponding Piper voice model (thorsten for German or ryan for English).
- The engine synthesizes the speech and generates a raw audio output.
- ffmpeg processes the audio into a Telegram-compatible OGG Opus format.
- The skill outputs the file path and a specific metadata tag that instructs the communication client to render the file as a voice bubble.
Voice Reply Setup
To integrate this capability into your Openclaw Skills, follow these installation steps:
- Install the sherpa-onnx runtime:
sudo mkdir -p /opt/sherpa-onnx
cd /opt/sherpa-onnx
curl -L -o sherpa.tar.bz2 "https://github.com/k2-fsa/sherpa-onnx/releases/download/v1.12.23/sherpa-onnx-v1.12.23-linux-x64-shared.tar.bz2"
sudo tar -xjf sherpa.tar.bz2 --strip-components=1
- Download the required voice models to
/opt/piper-voicesand install ffmpeg:
sudo apt install -y ffmpeg
- Configure the environment variables to point to your installation directories:
export SHERPA_ONNX_DIR="/opt/sherpa-onnx"
export PIPER_VOICES_DIR="/opt/piper-voices"
Voice Reply Data Schema & Taxonomy
The skill manages audio assets and configuration using the following structure:
| Type | Format | Description |
|---|---|---|
| Input | String | The text content to be converted to speech |
| Output Tag | [[audio_as_voice]] |
A directive for Telegram to display the file as a voice note |
| Output Path | MEDIA:/tmp/voice-reply-output.ogg |
The absolute path to the generated Opus audio file |
| Voice Models | .onnx / tokens.txt |
Local model files stored in the Piper voices directory |
Voice Reply Advanced Features
- Automatic language detection that intelligently switches between German and English voices based on text content.
- Telegram-native integration using the audio_as_voice tag for a professional chat experience.
- Support for custom voice expansion, allowing users to add any compatible Piper voice model from the Rhasspy library.
- Low-resource execution designed to run efficiently on local Linux servers without GPU acceleration.
SKILL.md
Loading
Related Openclaw Skills
Market Oracle
A professional financial analysis agent that provides real-time market data tracking and structured three-layer impact predictions for global economic events.
Ralph Loop Agent
An agent-based skill enabling iterative task execution and sophisticated automation loops within the OpenClaw environment.
PromptShield
A high-performance prompt injection firewall that protects AI agents from 113 threat patterns using multi-layered heuristic scoring.
CEO Notify Agents
An inter-agent messaging tool that allows a lead agent to broadcast notifications to sub-agents through shared memory files.
Clawearn
A modular prediction market trading platform for OpenClaw bots to automate Polymarket trades and wallet management.
Drip Director
A deterministic 8-stage image production pipeline designed for high-fidelity streetwear and fashion content creation.
Featured*
