voice-call
Start voice calls via the OpenClaw voice-call plugin.
Security Assessment
About voice-call
The voice-call skill enables AI agents and developers to programmatically initiate, manage, and monitor voice calls through the OpenClaw platform. It provides a unified interface for integrating with multiple telephony providers including Twilio, Telnyx, and Plivo, abstracting away provider-specific complexity. This skill solves the problem of building voice communication capabilities into AI applications without managing individual provider SDKs or handling different API patterns.
The skill offers comprehensive call lifecycle management through both CLI commands and programmatic tool actions. Core capabilities include initiating outbound calls with custom messages, continuing existing conversations, speaking to users mid-call, gracefully ending calls, and checking call status. The system supports multiple telephony providers with simple configuration switching, and includes a mock provider for development and testing without incurring network costs or requiring live credentials.
This skill is designed for developers building voice-enabled AI agents, customer service automation systems, notification platforms, or any application requiring programmable voice communication. It's particularly valuable for teams that need flexibility to switch between telephony providers or want to prototype voice features in development environments before deploying to production infrastructure.
FAQ
Which telephony providers are supported?
The voice-call skill supports Twilio, Telnyx, and Plivo as production providers. It also includes a mock provider for development and testing that simulates calls without making real network requests or incurring costs.
What configuration is required to use this skill?
You must enable the voice-call plugin in your OpenClaw configuration under plugins.entries.voice-call.enabled. Each provider requires specific credentials: Twilio needs accountSid and authToken, Telnyx needs apiKey and connectionId, and Plivo needs authId and authToken. All providers also require a fromNumber to originate calls.
Can I use this skill for both CLI and programmatic access?
Yes, the skill provides both CLI commands (openclaw voicecall) for manual operations and a voice_call tool with actions like initiate_call, continue_call, speak_to_user, end_call, and get_status for agent-initiated and programmatic call management.
How do I test voice call functionality without using real phone numbers?
Set the provider configuration to 'mock' in your plugin settings. The mock provider simulates call behavior without making actual network requests, allowing you to develop and test voice call logic without telephony costs or live credentials.
What call management actions are available?
The skill provides five core actions: initiate_call to start new calls with a message and optional recipient and mode, continue_call to extend ongoing conversations, speak_to_user to send messages during active calls, end_call to terminate calls gracefully, and get_status to retrieve current call state information.
Install voice-call
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
steipete/clawdis