smith-ai 1.0.0

Production AI Agent Framework - NVIDIA Nemotron, Voice, Canvas, Browser Sandbox, Multi-Channel Support

SmithAI

Production-ready Python AI Agent Framework with NVIDIA Nemotron, Voice, Canvas, Browser Sandbox, and Multi-Channel Support.

Features

• NVIDIA Nemotron - Advanced reasoning with LangChain integration
• Voice - Text-to-Speech and Speech-to-Text
• Canvas - Interactive code workspace with Python/JS execution
• Browser Sandbox - Playwright-based web automation
• Multi-Channel - WhatsApp, Signal, IRC, Matrix, Nostr, and more
• Memory - Persistent storage with MEMORY.md-style interface
• Skills - Markdown-based skill registry system
• Gateway - Message routing and session management
• TUI - Terminal user interface included

Installation

pip install smith-ai

Quick Start

from open_agent import create_deep_agent

agent = create_deep_agent(name="assistant")
await agent.initialize()
response = await agent.process_message("Hello!")
print(response)

Environment Variables

export NVIDIA_API_KEY=your_nvidia_api_key

Optional Dependencies

# Voice support
pip install smith-ai[voice]

# Browser automation
pip install smith-ai[browser]

# All extras
pip install smith-ai[all]

Core Modules

Module	Description
agents	DeepAgent runtime with tool orchestration
canvas	Interactive code execution workspace
channels	Multi-platform messaging adapters
gateway	Routing, sessions, and auth
memory	Persistent knowledge storage
sandbox	Browser automation
voice	TTS/STT capabilities

License

MIT License - Himan D [e-mail]