An interactive AI voice agent that can capture and transcribe speech in real-time, generate intelligent responses using the DeepSeek R1 (7B model) AI, and convert the responses back to natural speech for immediate playback. The agent maintains conversation context and supports cross-platform usage on macOS, Linux, and Windows.

DeepSeek R1 AI Voice Agent

A real-time AI voice assistant powered by DeepSeek R1 that enables seamless voice conversations through speech-to-text transcription, AI response generation, and text-to-speech synthesis.

🌟 Overview

This project creates an interactive AI voice agent that:

• Captures and transcribes speech in real-time using AssemblyAI
• Generates intelligent responses using DeepSeek R1 (7B model) via Ollama
• Converts AI responses back to natural speech using ElevenLabs
• Streams audio responses for immediate playback

✨ Features

• Real-time Speech Recognition: High-quality speech-to-text transcription with AssemblyAI
• Advanced AI Responses: Powered by DeepSeek R1's reasoning capabilities
• Natural Voice Synthesis: Professional text-to-speech with ElevenLabs
• Streaming Audio Playback: Low-latency audio streaming for responsive conversations
• Conversation Memory: Maintains context throughout the conversation
• Cross-platform Support: Works on macOS, Linux, and Windows

🔧 Prerequisites

API Keys Required

• AssemblyAI API Key: Get your free API key
• ElevenLabs API Key: Sign up for ElevenLabs

System Dependencies

Install Ollama

Download and install Ollama from ollama.com

Install PortAudio

Ubuntu/Debian:

sudo apt update && sudo apt install portaudio19-dev

macOS:

brew install portaudio

Windows: PortAudio is typically included with the Python package installation.

Install MPV (macOS only)

brew install mpv

📦 Installation

1. Clone the Repository

git clone https://github.com/danieladdisonorg/DeepSeek-R1-Voice-Agent.git
cd DeepSeek-R1-Voice-Agent

2. Install Python Dependencies

pip install "assemblyai[extras]" ollama elevenlabs

3. Download DeepSeek R1 Model