A local-to-cloud bridge for processing Speech-to-Text with Whisper and Text-to-Speech via AWS Polly.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install voice-agent
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-agent 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 AI Voice Agent Backend?
The AI Voice Agent Backend provides a specialized client-side interface for local Speech-to-Text (STT) and cloud-based Text-to-Speech (TTS). By leveraging local Whisper models for transcription and AWS Polly for natural voice generation, this skill allows agents to communicate through audio channels without manual text conversion. It operates as a client interface connecting to a dedicated local API, ensuring low-latency processing and high-quality audio interactions.
This tool is specifically designed for developers building hands-free or voice-first Openclaw Skills, prioritizing audio-first communication. It removes the need for text-based commentary during voice interactions, allowing for a seamless conversational flow where the agent processes audio inputs and responds with generated speech files automatically.
AI Voice Agent Backend Use Cases
- Processing voice commands or audio messages sent by users directly within an agent workflow.
- Generating high-quality spoken responses for accessibility or interactive voice assistants.
- Transcribing meeting recordings or short audio notes for automated agent analysis.
- Automating voice-driven feedback loops in multimodal AI applications powered by Openclaw Skills.
How AI Voice Agent Backend Works
- The user provides an audio file to the agent environment.
- The agent invokes the transcribe tool to process the file using a local Whisper STT implementation.
- The resulting text is interpreted by the agent's logic to formulate a contextually relevant response.
- The agent uses the synthesize tool to convert its internal text response into an MP3 or OGG file via AWS Polly.
- The agent delivers the generated audio file back to the user as the primary response without additional text clutter.
AI Voice Agent Backend Setup
To utilize this within your Openclaw Skills environment, ensure the backend API is running at http://localhost:8000. Follow these steps to verify and use the tools:
# Check if the voice agent API is running and healthy
python3 scripts/client.py health
# Transcribe an audio file using local Whisper
python3 scripts/client.py transcribe "/path/to/audio/file.ogg"
# Generate audio from text using AWS Polly
python3 scripts/client.py synthesize "Text to speak" --output "/path/to/output.mp3"
AI Voice Agent Backend Data Schema & Taxonomy
| Data Type | Description | Format |
|---|---|---|
| Transcription Output | Plain text resulting from the Whisper STT process. | String |
| Synthesis Input | The text response generated by the AI agent. | String |
| Synthesis Output | The final audio file returned to the user. | .mp3 or .ogg |
| Service Health | Status check of the connection to the local Voice Agent API. | Boolean/JSON |
AI Voice Agent Backend Advanced Features
- Local Whisper STT: Ensures privacy and low-latency transcription by processing audio locally.
- AWS Polly Integration: Supports professional-grade, natural-sounding speech synthesis for agent responses.
- Silent Delivery Protocol: Automatically suppresses text commentary when audio files are generated to maintain voice-first immersion.
- Backend Health Monitoring: Built-in diagnostic tools to ensure the local API server is responsive before attempting tasks.
- Multi-Agent Compatibility: Can be shared across different Openclaw Skills to provide a unified voice for complex agent teams.
SKILL.md
Loading
Related Openclaw Skills
Garmin Tracker
A specialized skill for synchronizing Garmin Connect activity history and training schedules into a structured JSON database.
bun-do-api
A local-first REST API for managing tasks, recurring payments, and project progress logs with full JSON persistence.
Microsoft Learn MCP
A bridge for AI agents to query, search, and retrieve live Microsoft technical documentation and official code samples.
Excalidraw Diagram Generator
An automated diagramming tool that converts text descriptions into hand-drawn Excalidraw visuals for Obsidian and web platforms.
Google Analytics
A comprehensive integration for managing Google Analytics 4 properties and retrieving reporting data via managed OAuth.
News Writing
A comprehensive journalistic workflow for AI agents to collect materials, verify facts, and generate publication-ready news content.
METADATA
Featured*
