MultiAgentes provides a modular architecture for defining environments and agents, supporting synchronous and asynchronous multi-agent interactions. It includes base classes for environments and agents, predefined scenarios for cooperative and competitive tasks, tools for customizing reward functions, and APIs for agent communication and observation sharing. Visualization utilities allow real-time monitoring of agent behaviors, while logging modules record performance metrics for analysis. The framework integrates seamlessly with Gym-compatible reinforcement learning libraries, enabling users to train agents using existing algorithms. MultiAgentes is designed for extensibility, allowing developers to add new environment templates, agent types, and communication protocols to suit diverse research and educational use cases.

MultiAgentSystems is designed to simplify the process of building and evaluating multi-agent reinforcement learning (MARL) applications. The platform includes implementations of state-of-the-art algorithms like MADDPG, QMIX, VDN, and centralized training with decentralized execution. It features modular environment wrappers compatible with OpenAI Gym, communication protocols for agent interaction, and logging utilities to track metrics such as reward shaping and convergence rates. Researchers can customize agent architectures, tune hyperparameters, and simulate settings including cooperative navigation, resource allocation, and adversarial games. With built-in support for PyTorch, GPU acceleration, and TensorBoard integration, MultiAgentSystems accelerates experimentation and benchmarking in collaborative and competitive multi-agent domains.

Shepherding is an open-source simulation framework designed for reinforcement learning researchers and developers to study and implement multi-agent herding tasks. It provides a Gym-compatible environment where agents can be trained to perform behaviors such as flanking, collecting, and dispersing target groups across continuous or discrete spaces. The framework includes modular reward shaping functions, environment parameterization, and logging utilities for monitoring training performance. Users can define obstacles, dynamic agent populations, and custom policies using TensorFlow or PyTorch. Visualization scripts generate trajectory plots and video recordings of agent interactions. Shepherding’s modular design allows seamless integration with existing RL libraries, enabling reproducible experiments, benchmarking of novel coordination strategies, and rapid prototyping of AI-driven herding solutions.

Text-to-Reward provides a pipeline to train reward models that map text-based task descriptions or feedback into scalar reward values for RL agents. Leveraging transformer-based architectures and fine-tuning on collected human preference data, the framework automatically learns to interpret natural language instructions as reward signals. Users can define arbitrary tasks via text prompts, train the model, and then incorporate the learned reward function into any RL algorithm. This approach eliminates manual reward shaping, boosts sample efficiency, and enables agents to follow complex multi-step instructions in simulated or real-world environments.