ClawCall AI for Openclaw

A professional AI voice skill that executes outbound phone calls with natural conversation flow, custom personalities, and full transcript recording.

shreyjindal81
v1.0.1
Feb 11, 2026
4
2.9k
0

Install & Download

1. ClawHub CLI

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

npx clawhub@latest install clawcall-ai

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 clawcall-ai 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 ClawCall AI?

ClawCall AI is a specialized communication tool designed for the Openclaw Skills ecosystem, enabling developers to automate real-world phone outreach. It leverages Telnyx for telephony infrastructure and Deepgram for high-performance voice processing, allowing AI agents to handle follow-ups, reminders, and complex task-oriented conversations with human-like fluidity.

Unlike simple robocallers, this skill prioritizes context and personality. By providing detailed background information, the agent can act as a specific employee or representative, ensuring a professional and empathetic interaction. It is ideal for bridging the gap between digital workflows and traditional voice communication channels.

ClawCall AI Use Cases

  • Scheduling and confirming appointments for clinics, salons, or professional services.
  • Following up on outstanding invoices or balances with specific payment instructions.
  • Conducting multi-stage interactions where an agent needs to call back with requested information.
  • Sending automated yet personalized reminders for post-operative care or service check-ups.

How ClawCall AI Works

  1. The user triggers the skill by providing a destination phone number (E.164 format) and a rich context string including personality and task details.
  2. The skill initiates a connection via Telnyx and establishes a real-time audio stream using a managed ngrok tunnel or public WebSocket.
  3. The AI agent processes speech-to-text via Deepgram and generates natural responses using ElevenLabs voice synthesis.
  4. During the call, the agent follows the specific task context, handling questions and providing details as instructed.
  5. Once the call terminates, the skill saves a local audio recording and returns a full text transcript for further processing within the Openclaw Skills environment.

ClawCall AI Setup

Before running the skill, ensure you have set the required environment variables: TELNYX_API_KEY, TELNYX_CONNECTION_ID, TELNYX_PHONE_NUMBER, and DEEPGRAM_API_KEY.

First, install the necessary dependencies:

npm install

To initiate a call using a managed tunnel, use the following command structure:

node telnyx_voice_agent.js --to "+15551234567" --ngrok --personality "A warm receptionist" --task "Follow up on the appointment for Max."

ClawCall AI Data Schema & Taxonomy

ClawCall AI organizes interaction data into transcripts and media files for easy auditing and record-keeping:

Data Type Description Location
Transcript Full text log of the conversation returned as output. Openclaw Logs
Audio Recording MP3/WAV file of the call session. ./recordings
Task Metadata The personality and task instructions provided during invocation. Command Arguments
Session Logs Connection and recording lifecycle events (e.g., URL discovery). CLI Output

ClawCall AI Advanced Features

  • Multi-turn continuity: Pass previous call transcripts into the task context for seamless follow-up conversations.
  • Custom Voice Selection: Choose between specific ElevenLabs voices like Rachel, Adam, or Josh to match your brand tone.
  • Automated Tunneling: Integrated ngrok support simplifies local testing without complex network configuration.
  • Privacy-First Recording: Automatically downloads recordings to local storage and deletes them from the cloud provider.
  • Context-Specific Logic: The agent is programmed to never assume data, ensuring accuracy and reliability during professional outreach.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Requires
Bins nodenpm
Github Stars: 0
forks: 0

Featured*