claude-phone  by theNetworkChuck

Voice interface for conversational AI agents

Created 2 months ago
257 stars

Top 98.2% on SourcePulse

GitHubView on GitHub
Project Summary

This project provides a voice interface for Claude Code, enabling users to call their AI and have the AI call them back via SIP/3CX. It targets developers and power users of Claude Code who wish to integrate conversational AI capabilities into phone systems, offering a novel way to interact with AI for tasks, status checks, or alerts.

How It Works

Claude Phone integrates with 3CX for call routing, acting as a SIP client. It leverages Claude Code CLI for AI processing, ElevenLabs for text-to-speech synthesis, and OpenAI Whisper for speech-to-text transcription. The system comprises a voice-app (typically Dockerized) that handles real-time communication protocols and an claude-api-server that interfaces with the Claude Code backend. This architecture allows for both inbound calls to interact with Claude and outbound calls initiated by the server for alerts or conversations.

Quick Start & Requirements

  • Installation: Run curl -sSL https://raw.githubusercontent.com/theNetworkChuck/claude-phone/main/install.sh | bash. The installer checks for and offers to install Node.js 18+, Docker, and Git.
  • Setup: Execute claude-phone setup for an interactive wizard to configure the deployment mode (Voice Server, API Server, or Both).
  • Run: Start services with claude-phone start.
  • Prerequisites:
    • 3CX Cloud Account (free tier available).
    • ElevenLabs API Key.
    • OpenAI API Key.
    • Claude Code CLI (requires Claude Max subscription, available at claude.ai/code).
    • Node.js 18+, Docker, Git.
  • Platform Support: macOS and Linux (including Raspberry Pi) are fully supported. Windows is not supported natively but may function with WSL.

Highlighted Details

  • Supports flexible deployment modes: All-in-One (single machine) or Split Mode (e.g., Raspberry Pi for voice, separate machine for API server).
  • Allows customization of individual SIP extensions with unique names, voices, and personality prompts.
  • Exposes RESTful API endpoints for programmatic control, including initiating outbound calls (/api/outbound-call) and querying call status.

Maintenance & Community

The README does not detail specific contributors, sponsorships, or community channels (like Discord/Slack). However, it includes a claude-phone update command for maintaining the software.

Licensing & Compatibility

The project is released under the MIT License, which is permissive and generally allows for commercial use and integration into closed-source projects.

Limitations & Caveats

Native support for Windows is absent. Issues such as audio connectivity problems may arise if the API server is unreachable, and extension registration can fail if the 3CX SBC is not operational.

Health Check
Last Commit

1 month ago

Responsiveness

Inactive

Pull Requests (30d)
4
Issues (30d)
1
Star History
67 stars in the last 30 days

Explore Similar Projects

Starred by Chip Huyen Chip Huyen(Author of "AI Engineering", "Designing Machine Learning Systems"), Vincent Weisser Vincent Weisser(Cofounder of Prime Intellect), and
12 more.

rasa by RasaHQ

0.0%
21k
AI framework for automating text and voice conversations
Created 9 years ago
Updated 1 month ago
Feedback? Help us improve.