v0.2.0

🎉 Major Update: Introducing Swarm Mode for Distributed AI Agent Execution! 🎉

This release marks a significant leap forward for SmolagentsCrew, introducing powerful new features for distributed task execution and collaborative AI agent workflows.

New Features:

🌐 Swarm Mode:

• Distributed Task Execution: Run your AI agent tasks across multiple machines or nodes, enabling parallel processing and high scalability.
• Automatic Load Balancing: Tasks are intelligently distributed across available nodes, optimizing resource utilization and throughput.
• Cross-Node Dependency Management: Maintain task dependencies across different nodes. The system ensures that tasks are executed in the correct order, even when distributed.
• Fault Tolerance: The swarm is designed to be resilient. It can handle node failures and retries, keeping your agent workflows running smoothly.

📊 Swarm Visualization:

• Text-Based Workflow Representation: Get a quick overview of the swarm's workflow with crew.print_swarm_flow().
• Interactive Graph Visualization: Use crew.visualize_swarm_flow() to generate a graphical representation of your swarm's task flow, including task dependencies, node assignments, and data flow.

🔑 Key Swarm Components:

• SwarmCrew: A new class that extends the core Crew functionality to manage distributed tasks.
• SwarmManager: Coordinates task execution across nodes and manages task dependencies in a distributed setting.
• SwarmNode: Represents a machine or node participating in the swarm, defining the agents it can execute.

Improvements:

• Enhanced Scalability: The introduction of swarm mode allows you to scale your AI agent operations to handle significantly larger and more complex tasks.
• Increased Robustness: The distributed nature of the swarm provides fault tolerance, making your AI workflows more reliable.
• Improved Visualization: The visualization options offer intuitive views of complex task flows.
• Improved documentation All the documentation was improved to explain the new features and how to use it.

Installation:

pip install smolagents-crew[viz] # includes visualization tools

None. This release is designed to be backward-compatible with existing code, but maybe you will have to adapt your code to use the new classes.

Full Changelog: https://github.com/antunsz/smolagents-crew/commits/v0.2.0

v0.1.1

What's Changed

• adding time tracking for tasks and advanced crew features by @antunsz in #1

New Contributors

• @antunsz made their first contribution in #1

Full Changelog: v0.1.0...v0.1.1

v0.1.0

Full Changelog: https://github.com/antunsz/smolagents-crew/commits/v0.1.0