A comprehensive integration that turns your Openclaw instance into a voice-activated phone assistant using ElevenLabs and Twilio.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install phone-voice
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 phone-voice 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 Phone Voice Integration?
The Phone Voice Integration is a sophisticated bridge that enables bidirectional voice communication with your AI agent over standard telephone lines. By leveraging the power of ElevenLabs Agents and Twilio, this addition to the collection of Openclaw Skills allows you to call your bot from any phone, authenticate via caller ID or PIN, and interact with a persistent memory context derived from your local environment. It effectively bridges the gap between high-quality neural text-to-speech and the logic of your local AI workspace.
This skill is designed for developers who need hands-free access to their automated workflows. Unlike basic voice tutorials, this implementation prioritizes security through caller ID whitelisting and cost management with granular tracking. By sitting between the ElevenLabs API and Anthropic Claude, the bridge server ensures that every phone call is personalized using your existing local memory files.
Phone Voice Integration Use Cases
- Accessing your AI agent and tools while driving or away from a computer.
- Implementing a secure voice-activated home or office receptionist for call screening.
- Remotely triggering tasks and setting reminders via natural language phone calls.
- Monitoring and managing the operational costs of voice-based AI interactions through automated logging.
How Phone Voice Integration Works
- A caller dials the designated Twilio phone number which is linked to an ElevenLabs Agent.
- Twilio routes the audio stream to ElevenLabs for real-time speech-to-text and low-latency response generation.
- The ElevenLabs Agent sends standardized chat completion requests to a custom FastAPI bridge server.
- The bridge server performs security checks (PIN or Whitelist) and injects local context from MEMORY.md and USER.md.
- The enriched prompt is sent to Anthropic Claude, and the response is streamed back through ElevenLabs TTS to the caller.
- Upon completion, the bridge logs the call duration and calculates the total cost across all providers.
Phone Voice Integration Setup
1. Configure the Bridge
Set up a FastAPI server to act as the intermediary. Create a .env file containing your ANTHROPIC_API_KEY, ELEVENLABS_API_KEY, and TWILIO credentials.
2. Expose Localhost
Use a Cloudflare Tunnel for a permanent, secure connection to your bridge:
# Install and login
brew install cloudflare/cloudflare/cloudflared
cloudflared tunnel login
# Create and run the tunnel
cloudflared tunnel create voice-bridge
cloudflared tunnel --url http://localhost:8013 run voice-bridge
3. ElevenLabs & Twilio Link
In the ElevenLabs dashboard, create a new Conversational AI agent. Set the Custom LLM URL to your tunnel address and provide your bridge authorization token. Finally, enter your Twilio Account SID and Auth Token in the Phone section to claim your number.
Phone Voice Integration Data Schema & Taxonomy
The skill manages interaction data and security through several structured files:
| File | Format | Purpose |
|---|---|---|
contacts.json |
JSON | A whitelist mapping phone numbers to names and roles for Caller ID auth. |
costs.jsonl |
JSONL | Append-only log tracking call SIDs, duration, and cost breakdown per provider. |
MEMORY.md |
Markdown | Long-term context and facts injected into the system prompt. |
USER.md |
Markdown | Personal profile data including location and timezone for live context. |
Phone Voice Integration Advanced Features
- Voice PIN Challenge: Adds an extra layer of security by requiring a spoken password for sensitive tasks or unrecognized numbers.
- Call Screening: Automatically switches to a receptionist mode for unknown callers to take messages without exposing tools.
- Office Hours Logic: Restricts the agent's availability based on specific timezones and business hours, routing off-hour calls to voicemail.
- Outbound API: Includes a dedicated endpoint to trigger outbound calls for reminders or alerts directly from the CLI or other Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
UI Skills
A rigorous set of architectural constraints and design principles that guide AI agents in building accessible, performant, and visually consistent user interfaces.
Tencent Map API Integration
A comprehensive integration of Tencent Map Web Services for advanced location search, route planning, and coordinate transformations.
Cross-Device Sync Skill
A synchronization tool that uses GitHub as a central repository to keep OpenClaw memory files consistent across all your devices.
Baidu Map Integration
Integrate comprehensive Baidu Map Web Service APIs for location searching, geocoding, and routing.
Taildrop
A command-line utility for retrieving files from the Tailscale Taildrop inbox to local Linux storage.
Douban Sync
A robust tool to export and incrementally sync your Douban library to local CSV files for personal archiving or Obsidian integration.
Featured*
