AI Call Center Backend

Overview: This project is the backend engine for a fully autonomous AI-powered call center. It integrates a large language model (LLM), speech recognition, and text-to-speech to manage real-time phone conversations via Asterisk.

The system is designed to orchestrate and stream AI responses in live phone calls using:

• Whisper for speech-to-text (STT)
• DeepSeek LLM for conversation logic
• Coqui TTS for natural-sounding audio synthesis
• Asterisk for SIP media and call control

Project Structure:

• orchestrator/ – Core Python modules handling LLM, TTS, STT, and Asterisk streaming logic
• scripts/ – Utility shell and Python scripts to start calls, launch LLM, and AGI interaction
• conf/ – Asterisk configuration
• .vscode/ – Development environment settings
• .gitignore – Standard Git exclusions

Key Components:

• orchestrator/main.py: Entry point for FastAPI app orchestrating AI call flow
• scripts/prepare_call.py: Initiates outbound calls via .call files
• scripts/agi_handler.sh: Asterisk AGI script to control call flow
• orchestrator/llm_streamer.py: Streams LLM tokens during conversation
• orchestrator/whisper_handler.py: Transcribes recorded voice to text
• orchestrator/tts_streamer.py: Converts LLM response into audio using Coqui TTS
• orchestrator/asterisk_streamer.py: Handles audio segment playback via Asterisk

Setup & Usage:

# Install Python dependencies (make sure virtualenv is active)
pip install -r requirements.txt

# Start the orchestrator API server
uvicorn orchestrator.main:app --host 0.0.0.0 --port 8100

# Start outbound call
python scripts/prepare_call.py +11234567890

Tech Stack:

• Python 3.10+
• FastAPI
• Whisper by OpenAI (STT)
• DeepSeek or Open LLM via vLLM (LLM)
• Coqui TTS
• Asterisk PBX

Author: Dewayne Johnson (dj@suncoast-software.com)