mindcraft-bots/mindcraft
Mindcraft
Mindcraft is a framework for connecting large language models to game clients to create autonomous characters that communicate and perform actions within a simulated environment. It functions as an orchestrator for bots, utilizing a system that bridges high-level AI instructions with low-level game protocol packets to enable the execution of in-game tasks.
The system uses retrieval-augmented generation to select relevant conversation history and code examples via embedding-based context retrieval. It supports the development of specific AI personas through profile configurations and facilitates multi-agent coordination, allowing multiple autonomous characters to collaborate through natural language and resource sharing to achieve shared objectives.
The framework includes a goal-oriented task system that decomposes complex objectives into procedural steps and validates them against world states. It provides capabilities for parallel world orchestration to run concurrent experiments across isolated environments, alongside text-to-speech integration for real-time audio narration of agent actions.
The project includes utilities for importing custom world files and synchronizing collected log files and configurations to remote cloud storage.
Features
- LLM-Driven Agent Loops - Provides the central loop that processes game state and LLM responses to execute actions in the world.
- LLM Integrations - Connects large language models to Minecraft game clients to create autonomous characters capable of performing complex in-game tasks.
- Multi-Agent Coordination Systems - Implements a framework where multiple specialized AI agents collaborate on complex in-game tasks by sharing state and delegating sub-processes.
- Agent Persona Configurations - Uses external configuration files to map backend models and system prompts to specific agent identities.
- Agentic RAG Platforms - Integrates reasoning agents with retrieval-augmented generation to fetch relevant conversation history and coding examples for improved behavior.
- Multi-Agent Collaboration Systems - Facilitates cooperation between multiple agents through natural language communication and resource sharing to achieve shared objectives.
- Agent Persona Definitions - Defines agent behavior and constraints through profile-based system prompts and model configurations.
- Autonomous Agent Simulations - Provides a simulated environment to evaluate the performance of autonomous AI agents on complex tasks.
- Autonomous Goal Pursuit - Executes predefined high-level goals such as item collection to drive autonomous agent behavior.
- LLM Model Integrations - Provides the mechanism for integrating large language models into the game environment to drive autonomous characters.
- Persona Behavioral Refinement - Refines character behaviors and prompts to control how AI agents interact and speak.
- Procedural Task Decomposition - Decomposes complex objectives into smaller procedural steps validated against the game world and inventory.
- Game Automation Bots - Uses the Mineflayer framework to automate in-game tasks like construction and item collection.
- Minecraft Bot Frameworks - Integrates large language models with Minecraft bots to create characters that communicate and act.
- Bot Orchestration Frameworks - Provides a management layer for Mineflayer bots to coordinate multi-agent collaboration and task execution via natural language.
- Contextual Information Retrieval - Uses embedding-based similarity search to select relevant conversation history and code snippets to ground AI responses.
- Contextual Retrieval for Agents - Implements embedding-based retrieval of conversation and code context to improve the precision of agent prompts.
- Game AI Environments - Provides a platform for running parallel isolated game worlds to evaluate how LLM agents handle procedural goals.
- World-State Goal Definitions - Allows creating new objectives by specifying target items and inventories to validate agent success.
- Game Protocol Abstractions - Uses Mineflayer to translate high-level AI instructions into low-level Minecraft protocol packets.
- Parallel World Simulations - Launches multiple isolated game instances simultaneously to run concurrent autonomous agent experiments.
- Scenario-Based Agent Evaluation - Runs agents through generated scenarios to evaluate their ability to complete complex procedural goals.
Star history
How this analysis was created: This summary and feature list are AI-generated from collected project material and can contain mistakes. Stars, license and language are imported from GitHub. Inclusion does not mean that we have tested or audited this project. Check the source documentation for any feature you depend on. Learn more on our About page.
