Multi-Agent-LLM is designed to streamline the orchestration of multiple AI agents powered by large language models. Users can define individual agents with unique personas, memory storage, and integrated external tools or APIs. A central AgentManager handles communication loops, allowing agents to exchange messages in a shared environment and collaboratively advance towards complex objectives. The framework supports swapping LLM providers (e.g., OpenAI, Hugging Face), flexible prompt templates, conversation histories, and step-by-step tool contexts. Developers benefit from built-in utilities for logging, error handling, and dynamic agent spawning, enabling scalable automation of multi-step workflows, research tasks, and decision-making pipelines.

DEf-MARL (Decentralized Execution Framework for Multi-Agent Reinforcement Learning) provides a robust infrastructure to execute and train cooperative agents without centralized controllers. It leverages peer-to-peer communication protocols to share policies and observations among agents, enabling coordination through local interactions. The framework integrates seamlessly with common RL toolkits like PyTorch and TensorFlow, offering customizable environment wrappers, distributed rollout collection, and gradient synchronization modules. Users can define agent-specific observation spaces, reward functions, and communication topologies. DEf-MARL supports dynamic agent addition and removal at runtime, fault-tolerant execution by replicating critical state across nodes, and adaptive communication scheduling to balance exploration and exploitation. It accelerates training by parallelizing environment simulations and reducing central bottlenecks, making it suitable for large-scale MARL research and industrial simulations.