AnyRobo - Create Your Own Robo Assistant

AnyRobo is an advanced speech-to-speech AI assistant framework that enables you to create your own real-life version of sci-fi AI assistants like JARVIS (from Iron Man) or GLADOS (from Portal). Powered by state-of-the-art machine learning models, AnyRobo listens to your voice, understands your requests, and responds with natural-sounding speech in real-time.

Why AnyRobo?

Have you ever wanted to create your own JARVIS or GLADOS? AnyRobo provides a fully modular framework that allows you to:

• Create voice-powered AI assistants with unique personalities
• Customize voice profiles to sound like your favorite AI characters
• Build advanced conversation capabilities with state-of-the-art language models
• Deploy your assistant on macOS with optimized performance for Apple Silicon

Core Technologies

• Speech Recognition: Whisper MLX - Optimized for Apple Silicon
• Language Understanding: Llama 3.2 - Advanced language model for contextual responses
• Voice Synthesis: Kokoro-82M - High-quality text-to-speech

Features

• Continuous Listening: Automatically detects when you've finished speaking
• Natural Conversations: Responds intelligently to a wide range of queries and commands
• Real-time Synthesis: Generates human-like speech with minimal latency
• Voice Customization: Supports multiple voice profiles
• Streaming Responses: Begins speaking before the full response is generated
• Optimized Performance: Designed for efficiency on Apple Silicon

Plan

• JARVIS assistant with UI
• Natural voice interaction
• Functional calling
• Web search
• MCP support
• Computer vision: See what's in front of the camera
• Build your own AI character
• 3D-printed JARVIS case
• GLADOS assistant with UI
• 3D-printed GLADOS case
• Physical button to trigger actions
• Physical robot bodies
• .... [add your ideas here]

Installation

Quick Install (from PyPI)

pip install anyrobo

Install from Source

git clone https://github.com/vietanhdev/anyrobo.git
cd anyrobo
pip install -e .

Setup Dependencies

AnyRobo requires Ollama for LLM support:

# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh
# Pull the required model
ollama pull llama3.2

Usage

• Test voice assistant:

python examples/test_assistant.py

Create Your Own AI Character

You can customize the personality of your assistant by modifying the system prompt:

# JARVIS from Iron Man
system_prompt = (
    "You are J.A.R.V.I.S., an advanced AI assistant. "
    "Respond with a mix of helpfulness, light sarcasm, and technical prowess."
)

# GLADOS from Portal
system_prompt = (
    "You are GLaDOS, an AI with a dark sense of humor. "
    "Respond to queries sarcastically, occasionally mentioning cake or testing."
)

# HAL 9000 from 2001: A Space Odyssey
system_prompt = (
    "You are HAL 9000. Be calm, logical, and slightly ominous in your responses. "
    "Speak in a slow, deliberate manner and be excessively literal."
)

Configuration Options

Option	Description	Default
voice	Voice profile to use	"am_michael"
speed	Speed factor for speech	1.2
silence_threshold	Volume level that counts as silence	0.02
silence_duration	Seconds of silence before cutting recording	1.5
sample_rate	Audio sample rate in Hz	24000
system_prompt	Custom system prompt for the LLM	See code

Troubleshooting

• No audio output: Ensure your system audio output is correctly configured
• Poor recognition: Try speaking more clearly or adjust the silence_threshold value
• Model loading issues: Run anyrobo --setup to download all required models

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Acknowledgements

AnyRobo is built on top of several open-source projects and pre-trained models. We're grateful to the developers and researchers who make their work available to the community.