DeepSeek R1 AI Voice Agent

A sophisticated real-time AI voice assistant leveraging DeepSeek R1's advanced reasoning capabilities for seamless conversational interactions through cutting-edge speech processing technology.

🎯 Project Overview

This enterprise-grade voice agent delivers an exceptional conversational AI experience by integrating:

• Advanced Speech Recognition: High-fidelity real-time transcription powered by AssemblyAI
• Intelligent Response Generation: DeepSeek R1's sophisticated reasoning and contextual understanding
• Natural Voice Synthesis: Professional-grade text-to-speech conversion via ElevenLabs
• Low-Latency Streaming: Optimized audio processing for responsive real-time interactions
• Contextual Memory: Persistent conversation context for coherent multi-turn dialogues
• Cross-Platform Compatibility: Seamless operation across macOS, Linux, and Windows environments

🚀 Core Capabilities

Real-Time Speech Processing

• High-Accuracy Transcription: Leverages AssemblyAI's state-of-the-art speech recognition
• Noise Reduction: Advanced audio preprocessing for optimal transcription quality
• Multi-Language Support: Extensible language processing capabilities

Intelligent AI Responses

• DeepSeek R1 Integration: Harnesses the power of DeepSeek's 7B parameter model
• Contextual Understanding: Maintains conversation flow and context awareness
• Reasoning Capabilities: Advanced logical reasoning and problem-solving abilities
• Response Optimization: Character-limited responses for optimal real-time performance

Professional Audio Output

• ElevenLabs Integration: Industry-leading text-to-speech synthesis
• Voice Customization: Configurable voice models and parameters
• Streaming Architecture: Low-latency audio streaming for immediate playback
• Audio Quality: High-fidelity 16kHz sample rate for crystal-clear output

📋 System Requirements

Essential Dependencies

API Services

• AssemblyAI Account: Register for free API access
• ElevenLabs Account: Create ElevenLabs account

Core Software Components

Ollama Installation

# Download and install from official source
curl -fsSL https://ollama.com/install.sh | sh

PortAudio System Library

Ubuntu/Debian Systems:

sudo apt update && sudo apt install portaudio19-dev

macOS Systems:

brew install portaudio

Windows Systems: PortAudio is automatically included with Python package installation.

MPV Media Player (macOS Only)

brew install mpv

🛠️ Installation Guide

Step 1: Repository Setup

git clone https://github.com/bigdata5911/DeepSeek-R1-Voice-Agent.git
cd DeepSeek-R1-Voice-Agent

Step 2: Python Environment Configuration

# Install required Python packages
pip install "assemblyai[extras]" ollama elevenlabs

Step 3: AI Model Deployment

# Download DeepSeek R1 model via Ollama
ollama pull deepseek-r1:7b

Step 4: API Configuration

Edit AIVoiceAgent.py and configure your API credentials:

# AssemblyAI Configuration
aai.settings.api_key = "YOUR_ASSEMBLYAI_API_KEY"

# ElevenLabs Configuration
self.client = ElevenLabs(api_key="YOUR_ELEVENLABS_API_KEY")

🎮 Usage Instructions

Launching the Voice Agent

python AIVoiceAgent.py

Interaction Workflow

1. Voice Input: Speak clearly into your microphone
2. Processing Pipeline: Speech → Transcription → AI Analysis → Response Generation
3. Audio Output: AI response converted to speech and streamed in real-time
4. Conversation Continuation: Context-aware dialogue progression

Session Management

• Start: Execute the Python script to initiate the voice agent
• Stop: Press Ctrl+C to gracefully terminate the session

⚙️ Advanced Configuration

Model Parameters

• AI Engine: DeepSeek R1 7B (configurable via Ollama)
• Voice Engine: ElevenLabs Turbo v2 (customizable)
• Response Constraints: 300-character limit for optimal performance
• Audio Specifications: 16kHz sample rate for professional quality

Customization Options

• System Prompt Modification: Adjust AI behavior and personality
• Response Length Tuning: Modify character limits for different use cases
• Voice Model Selection: Choose from ElevenLabs voice library
• Audio Stream Optimization: Fine-tune streaming parameters

🔧 Troubleshooting Guide

Common Issues and Solutions

Module Import Errors

# Reinstall AssemblyAI with extras
pip install "assemblyai[extras]"

Ollama Connection Issues

# Verify Ollama service status
ollama serve

# Check model availability
ollama list

Audio Device Problems

• Verify microphone permissions and access
• Confirm PortAudio installation integrity
• Test audio input with system applications

API Service Errors

• Validate API key authenticity and permissions
• Monitor API usage quotas and limits
• Ensure stable network connectivity

Performance Optimization

• Hardware Recommendations: Use high-quality microphone for optimal transcription
• Network Requirements: Stable broadband connection for API services
• System Resources: Close unnecessary applications to maximize performance

🏗️ Technical Architecture

┌─────────────────┐    ┌──────────────┐    ┌─────────────────┐
│   Audio Input   │───▶│  AssemblyAI  │───▶│   DeepSeek R1   │
│   (Microphone)  │    │ (Speech-to-  │    │ (AI Response    │
└─────────────────┘    │  Text)       │    │  Generation)    │
                       └──────────────┘    └─────────────────┘
                                                      │
┌─────────────────┐    ┌──────────────┐              │
│   Audio Output  │◀───│  ElevenLabs  │◀─────────────┘
│   (Speakers)    │    │ (Text-to-    │
└─────────────────┘    │  Speech)     │
                       └──────────────┘

📄 License Information

This project is released under open source licensing. Please refer to the repository for complete license details.

🤝 Contribution Guidelines

We welcome contributions from the community! Please consider:

• Issue Reporting: Submit detailed bug reports and feature requests
• Code Contributions: Pull requests for improvements and new features
• Documentation: Help improve project documentation and guides
• Testing: Assist with testing and quality assurance

📞 Support Resources

For technical assistance and community support:

• GitHub Issues: Open detailed issues for bug reports and feature requests
• Troubleshooting: Refer to the comprehensive troubleshooting section above
• API Documentation: Consult official documentation for AssemblyAI, Ollama, and ElevenLabs
• Community: Engage with the open-source community for best practices

---

Technical Note: This application requires persistent internet connectivity for API service integration and adequate system resources for local DeepSeek R1 model execution via Ollama.

Developer: @bigdata5911