This project demonstrates a comprehensive framework for creating retrieval-augmented AI agents using LlamaIndex. It guides developers through the entire workflow, starting with document ingestion and vector store creation, followed by defining a custom agent loop for contextual question-answering. Leveraging LlamaIndex's powerful indexing and retrieval capabilities, users can integrate any OpenAI-compatible language model, customize prompt templates, and manage conversation flows via a CLI interface. The modular architecture supports various data connectors, plugin extensions, and dynamic response customization, enabling rapid prototyping of enterprise-grade knowledge assistants, interactive chatbots, and research tools. This solution streamlines building domain-specific AI agents in Python, ensuring scalability, flexibility, and ease of integration.

Multi-Agent-RAG provides a modular framework for constructing retrieval-augmented generation (RAG) applications by orchestrating multiple specialized AI agents. Developers configure individual agents: a retrieval agent connects to vector stores to fetch relevant documents; a reasoning agent performs chain-of-thought analysis; and a generation agent synthesizes final responses using large language models. The framework supports plugin extensions, configurable prompts, and comprehensive logging, enabling seamless integration with popular LLM APIs and vector databases to improve RAG accuracy, scalability, and development efficiency.

ChainLite streamlines creation of AI agents by abstracting the complexities of LLM orchestration into reusable chain modules. Using simple Python decorators and configuration files, developers define agent behaviors, tool interfaces and memory structures. The framework integrates with popular LLM providers (OpenAI, Cohere, Hugging Face) and external data sources (APIs, databases), allowing agents to fetch real-time information. With a built-in browser-based UI powered by Streamlit, users can inspect token-level conversation history, debug prompts, and visualize chain execution graphs. ChainLite supports multiple deployment targets, from local development to production containers, enabling seamless collaboration between data scientists, engineers, and product teams.

Minerva is an extensible AI agent framework designed to automate complex workflows using large language models. Developers can integrate external tools—such as web search, API calls, or file processors—define custom planning strategies, and manage conversational or persistent memory. Minerva supports both synchronous and asynchronous task execution, configurable logging, and a plugin architecture, making it easy to prototype, test, and deploy intelligent agents capable of reasoning, planning, and tool use in real-world scenarios.

Pacman AI offers a fully functional Python-based environment and agent framework for the classic Pacman game. The project implements key reinforcement learning algorithms—Q-learning and value iteration—to allow the agent to learn optimal policies for pill collection, maze navigation, and ghost avoidance. Users can define custom reward functions and adjust hyperparameters such as learning rate, discount factor, and exploration strategy. The framework supports metric logging, performance visualization, and reproducible experiment setups. It is designed for easy extension, letting researchers and students integrate new algorithms or neural network-based learning approaches and benchmark them against baseline grid-based methods within the Pacman domain.

Real-Agents is designed to simplify the creation and orchestration of AI-powered agents that can perform complex tasks autonomously. Built on Python and compatible with major large language models, the framework features a modular design comprising core components for language understanding, reasoning, memory storage, and tool execution. Developers can rapidly integrate external services like web APIs, databases, and custom functions to extend agent capabilities. Real-Agents supports memory mechanisms to retain context across interactions, enabling multi-turn conversations and long-running workflows. The platform also includes utilities for logging, debugging, and scaling agents in production environments. By abstracting low-level details, Real-Agents streamlines the development cycle, allowing teams to focus on task-specific logic and deliver powerful automated solutions.

TexasHoldemAgent provides a modular environment built on Python to train, evaluate, and deploy an AI-powered poker player for heads-up limit Texas Hold’em. It integrates a custom simulation engine with deep reinforcement learning algorithms, including DQN, for iterative policy improvement. Key capabilities include hand state encoding, action space definition (fold, call, raise), reward shaping, and real-time decision evaluation. Users can customize learning parameters, leverage CPU/GPU acceleration, monitor training progress, and load or save trained models. The framework supports batch simulation to test various strategies, generate performance metrics, and visualize win rates, empowering researchers, developers, and poker enthusiasts to experiment with AI-driven gameplay strategies.