anyrobo 0.2.1

A local-first voice AI assistant framework. Create your own JARVIS.

anyrobo

Build your own JARVIS — local-first voice AI

A local-first voice AI assistant framework. Create your own JARVIS with pluggable TTS, STT, and LLM backends. Works completely offline with local models.

Built by Viet-Anh Nguyen at NRL.ai.

Installation

pip install anyrobo

With all optional backends:

pip install anyrobo[all]

Quick Start

import anyrobo

# Create an assistant with a personality
assistant = anyrobo.Assistant(personality="jarvis")

# Add custom tools
def get_weather(location: str) -> str:
    """Get the current weather for a location."""
    return f"Sunny in {location}"

assistant.add_tool(get_weather, description="Get weather for a location")

# Chat with the assistant
response = assistant.chat("What's the weather in Paris?")
print(response)

# Stream responses
for chunk in assistant.chat("Tell me a story", stream=True):
    print(chunk, end="")

Personalities

jarvis = anyrobo.Personality.builtin("jarvis")   # Helpful butler
glados = anyrobo.Personality.builtin("glados")   # Sarcastic AI
default = anyrobo.Personality.builtin("assistant")  # Neutral assistant

Memory

memory = anyrobo.ConversationMemory(max_messages=100, persistence_path="memory.json")
memory.add("user", "Hello")
memory.add("assistant", "Hi there!")
results = memory.search("hello")

Events

@assistant.on("tool_call")
def log_tool(name, args):
    print(f"Calling {name} with {args}")

Configuration

Create a YAML config file:

personality: jarvis
llm:
  backend: ollama
  model: llama3
stt:
  backend: whisper
  model_size: base
tts:
  backend: pyttsx3

config = anyrobo.AssistantConfig.from_yaml("config.yaml")
assistant = anyrobo.Assistant(config=config)

CLI

anyrobo chat          # Interactive chat
anyrobo listen        # Voice input mode
anyrobo config show   # Show current config

Backends

LLM

• Ollama (default) - Local LLM inference
• OpenAI - OpenAI API

Speech-to-Text

• Whisper (default) - OpenAI Whisper local model
• Vosk - Offline STT

Text-to-Speech

• pyttsx3 (default) - Offline TTS
• ElevenLabs - Cloud TTS API

License

MIT License - see LICENSE for details.

Author

Viet-Anh Nguyen (@vietanhdev)