Jump to Content

📞General Disarray

🤖 Voice-powered AI assistant for SIP phone systems by CHAOS.CORP

🤖
ROBO CODED — This documentation was made with AI and may not be 100% sane. But the code does work! 🎉

📞General Disarray - SIP AI Assistant

A voice-powered AI assistant that answers phone calls, understands natural language, and can perform actions like setting timers, checking weather, scheduling callbacks, and more.

✨ Features

Feature	Description
🎙️ Voice Conversations	Natural speech-to-text and text-to-speech powered by Whisper and Kokoro
🤖 LLM Integration	Connects to OpenAI, vLLM, Ollama, LM Studio, and more
🔧 Built-in Tools	Weather, timers, callbacks, date/time, calculator
🔌 Plugin System	Easily add custom tools with Python
🌐 REST API	Initiate outbound calls, execute tools, schedule calls
🔗 Webhooks	Trigger calls from Home Assistant, n8n, and more
⏰ Scheduled Calls	One-time or recurring calls (daily briefings, reminders)
🗣️ Custom Phrases	Customize greetings, goodbyes, and responses via config
📊 Observability	Prometheus metrics, OpenTelemetry tracing, JSON logs

🚀 Quick Example

Call the assistant and say:

🗣️
"What's the weather like?"

sequenceDiagram
    participant User as 👤 User
    participant Agent as 🤖 SIP Agent
    participant STT as 🎤 Speaches
    participant LLM as 🧠 LLM
    participant Tool as 🌤️ Weather Tool
    
    User->>Agent: "What's the weather like?"
    Agent->>STT: Audio stream
    STT-->>Agent: Transcribed text
    Agent->>LLM: User query + context
    LLM-->>Agent: [TOOL:WEATHER]
    Agent->>Tool: Execute
    Tool-->>Agent: Weather data
    Agent->>LLM: Tool result
    LLM-->>Agent: Natural response
    Agent->>STT: Text to speech
    STT-->>Agent: Audio
    Agent->>User: "At Storm Lake, it's 44°..."

Assistant responds:

🤖
"At Storm Lake, as of 9:30 pm, it's 44 degrees with foggy conditions. Wind is calm."

💡 Use Cases

Use Case	Example
⏲️ Timers & Reminders	"Set a timer for 10 minutes"
📞 Callbacks	"Call me back in an hour"
🌤️ Weather Briefings	Scheduled morning weather calls
📅 Appointment Reminders	Outbound calls with confirmation
🚨 Alerts & Notifications	Webhook-triggered phone calls
🏠 Smart Home	Voice control via phone

🧠 Recommended Models

Quick reference for GPU-specific configurations. See Configuration for full details.

GPU	VRAM	Recommended LLM	STT Model
H100 / A100	80GB	`meta-llama/Llama-3.1-70B-Instruct`	`faster-whisper-large-v3`
DGX Spark	128GB	`meta-llama/Llama-3.1-70B-Instruct`	`faster-whisper-large-v3`
RTX 5090	32GB	`Qwen/Qwen2.5-32B-Instruct`	`faster-whisper-large-v3`
RTX 4090	24GB	`Qwen/Qwen2.5-14B-Instruct`	`faster-whisper-large-v3`
RTX 3090	24GB	`meta-llama/Llama-3.1-8B-Instruct`	`faster-whisper-medium`
RTX 4080	16GB	`meta-llama/Llama-3.1-8B-Instruct`	`faster-whisper-medium`
RTX 3080	10GB	`Qwen/Qwen2.5-7B-Instruct`	`faster-whisper-small`

🎬 Demo

# Example call flow
┌──────────────────────────────────────────────────────────────┐
│  📞 Incoming call from: +1 (555) 123-4567                    │
├──────────────────────────────────────────────────────────────┤
│  🤖 "Hello! This is your AI assistant. How can I help?"     │
│  👤 "What time is it?"                                       │
│  🤖 "It's 3:45 PM on Saturday, November 30th."              │
│  👤 "Set a timer for 5 minutes"                              │
│  🤖 "Timer set for 5 minutes. I'll let you know!"           │
│  👤 "Thanks, goodbye"                                        │
│  🤖 "Goodbye! Have a great day!"                            │
│  📴 Call ended (duration: 0:32)                              │
└──────────────────────────────────────────────────────────────┘

📚 Documentation

🚀 Getting Started — Installation & setup
⚙️ Configuration — Environment variables
🌐 API Reference — REST API endpoints
🔧 Built-in Tools — Available capabilities
🔌 Creating Plugins — Add custom tools
📖 Examples — Integration patterns

📦 Quick Install

# Clone the repository
git clone https://github.com/your-org/sip-agent.git
cd sip-agent

# Configure environment
cp .env.example .env
nano .env

# Start with Docker Compose
docker compose up -d

# Verify it's running
curl http://localhost:8080/health

Expected output:

{
  "status": "healthy",
  "sip_registered": true,
  "active_calls": 0
}

🆘 Support

Updated about 1 month ago