🤖 smolagents-rs

This is a rust implementation of HF smolagents library. It provides a powerful autonomous agent framework written in Rust that solves complex tasks using tools and LLM models.

✨ Features

• 🧠 Function-Calling Agent Architecture: Implements the ReAct framework for advanced reasoning and action.
• 🔍 Built-in Tools:
  • Google Search
  • DuckDuckGo Search
  • Website Visit & Scraping
• 🤝 OpenAI Integration: Works seamlessly with GPT models.
• 🎯 Task Execution: Enables autonomous completion of complex tasks.
• 🔄 State Management: Maintains persistent state across steps.
• 📊 Beautiful Logging: Offers colored terminal output for easy debugging.

✅ Feature Checklist

Models

• OpenAI Models (e.g., GPT-4, GPT-4 Turbo)
• Ollama Integration
• Hugging Face API support
• Open-source model integration via Candle
• Light LLM integration

Agents

• Tool-Calling Agent
• CodeAgent
• Planning Agent

Tools

• Google Search Tool
• DuckDuckGo Tool
• Website Visit & Scraping Tool
• RAG Tool
• More tools to come...

Other

• Sandbox environment
• Streaming output
• Improve logging
• Parallel execution

🚀 Quick Start

Warning: Since there is no implementation of a Sandbox environment, be careful with the tools you use. Preferrably run the agent in a controlled environment using a Docker container.

Using Docker

# Pull the image
docker pull your-username/smolagents-rs:latest

# Run with your OpenAI API key
docker run -e OPENAI_API_KEY=your-key-here smolagents-rs -t "What is the latest news about Rust programming?"

Building from Source

# Clone the repository
git clone https://github.com/yourusername/smolagents-rs.git
cd smolagents-rs

# Build the project
cargo build --release --features cli-deps

# Run the agent
OPENAI_API_KEY=your-key-here ./target/release/smolagents-rs -t "Your task here"

🛠️ Usage

smolagents-rs [OPTIONS] -t TASK

Options:
  -t, --task <TASK>          The task to execute
  -a, --agent-type <TYPE>    Agent type [default: function-calling]
  -l, --tools <TOOLS>        Comma-separated list of tools [default: duckduckgo,visit-website]
  -m, --model <TYPE>         Model type [default: open-ai]
  -k, --api-key <KEY>        OpenAI API key (only required for OpenAI model)
  --model-id <ID>            Model ID (e.g., "gpt-4" for OpenAI or "qwen2.5" for Ollama) [default: gpt-4o-mini]
  -u, --ollama-url <URL>     Ollama server URL [default: http://localhost:11434]
  -s, --stream               Enable streaming output
  -h, --help                 Print help

🌟 Examples

# Simple search task
smolagents-rs -t "What are the main features of Rust 1.75?"

# Research with multiple tools
smolagents-rs -t "Compare Rust and Go performance" -l duckduckgo,google-search,visit-website

# Stream output for real-time updates
smolagents-rs -t "Analyze the latest crypto trends" -s

🔧 Configuration

Environment Variables

• OPENAI_API_KEY: Your OpenAI API key (required).
• SERPAPI_API_KEY: Google Search API key (optional).

🏗️ Architecture

The project follows a modular architecture with the following components:

1. Agent System: Implements the ReAct framework.

2. Tool System: An extensible tool framework for seamless integration of new tools.

3. Model Integration: Robust OpenAI API integration for powerful LLM capabilities.

🚀 Why port to Rust?

Rust provides critical advantages that make it the ideal choice for smolagents-rs:

1. ⚡ High Performance:
   Zero-cost abstractions and no garbage collector overhead enable smolagents-rs to handle complex agent tasks with near-native performance. This is crucial for running multiple agents and processing large amounts of data efficiently.

2. 🛡️ Memory Safety & Security:
   Rust's compile-time guarantees prevent memory-related vulnerabilities and data races - essential for an agent system that handles sensitive API keys and interacts with external resources. The ownership model ensures thread-safe concurrent operations without runtime overhead.

3. 🔄 Powerful Concurrency:
   Fearless concurrency through the ownership system enable smolagents-rs to efficiently manage multiple agents and tools in parallel, maximizing resource utilization.

4. 💻 Universal Deployment:
   Compile once, run anywhere - from high-performance servers to WebAssembly in browsers. This allows smolagents-rs to run natively on any platform or be embedded in web applications with near-native performance.

Apart from this, its essential to push new technologies around agentic systems to the Rust ecoystem and this library aims to do so.

🤝 Contributing

Contributions are welcome! To contribute:

1. Fork the repository.
2. Create your feature branch (git checkout -b feature/amazing-feature).
3. Commit your changes (git commit -m 'Add some amazing feature').
4. Push to the branch (git push origin feature/amazing-feature).
5. Open a Pull Request.

⭐ Show Your Support

Give a ⭐️ if this project helps you or inspires your work!