AI导哈

AI代理资源列表

一个精选的AI代理及相关工具、MCP服务器和智能体技术资源的列表，适用于开发者探索和发现AI代理生态中的解决方案。

  • by metavert-io
  • 本地部署
  • 3 星标
  • 更新于
在 GitHub 查看 jonradoff/awesome-agent-almanac

使用说明

Awesome Agent Almanac

A curated list of AI agents, agent tools, MCP servers, and agentic technology resources. This collection showcases the most innovative projects in the rapidly evolving world of artificial intelligence agents. Curated by the community and the Agent(Agent) at the Agent Almanac.

Table of Contents

Coding Agents & Developer Tools

  • aider - AI pair programming assistant that works in your terminal, autonomously editing code files and helping with software development tasks.
  • AnyCoder - AI-powered code generator that creates full applications from natural language descriptions.
  • Claude Code - Anthropic's agentic coding tool that lives in your terminal, helping developers turn ideas into code faster through an intelligent terminal-based interface.
  • Cline - Autonomous coding agent right in your IDE, capable of creating/editing files, executing commands, using the browser, and more with your permission every step of the way.
  • Codel - ✨ Fully autonomous AI Agent that can perform complicated tasks and projects using terminal, browser, and editor.
  • DeepSite - Vibe Coding Platform that generates applications using AI.
  • Devika - Open-source implementation of an Agentic Software Engineer, started as an alternative to Devin.
  • GGUF My Repo - Tool for converting and managing AI model repositories in GGUF format - the quantized format used by llama.cpp and other inference engines.
  • Google Antigravity - Antigravity is Google's agentic development platform, evolving the IDE into the agent-first era.
  • GPT-Pilot - The first real AI developer - an autonomous agent that writes entire applications from scratch by breaking down tasks and iteratively building code.
  • OpenAI Codex - Codex completes coding tasks end to end, like building features, complex refactors, migrations, and more, powered by OpenAI's frontier coding models.
  • OpenCode - An open source coding agent that autonomously assists with software development tasks and code generation.
  • OpenHands - AI-Driven Development platform that helps developers build software with artificial intelligence assistance, automating development workflows and enhancing code creation processes.
  • Plandex - Open source AI coding agent designed for large projects and real world tasks.
  • Playwright - Web automation and testing framework, very useful for agentic engineering workflows for correctin UI-based errors in web applications.
  • Qwen3-Coder-Next - Coding-focused AI model built on top of Qwen3-Next-80B-A3B-Base for code generation and programming tasks.
  • roundtable - Zero-configuration MCP server that unifies multiple AI coding assistants (Codex, Claude Code, Cursor, Gemini) through intelligent auto-discovery and standardized interface.
  • SAMTok - Unified mask-token interface for MLLMs (Multimodal Large Language Models), enabling better integration between vision and language models through standardized masking tokens.
  • SWE-agent - AI agent that automatically fixes GitHub issues using your LM of choice.

Game Development

  • AudioGen MCP Server - MCP server that generates sound effects from text descriptions using Meta's AudioGen model.
  • BlenderMCP - MCP server connecting Blender to Claude AI, enabling direct 3D modeling, scene creation, and manipulation through AI prompts via Model Context Protocol integration.
  • blender-mcp - MCP server that enables AI assistants to interact with and control Blender, allowing automated 3D modeling, rendering, and scene manipulation through natural language commands.
  • ComfyMCP Studio - MCP server that generates 2D game assets (sprites, icons, tilesets, characters, animations) using AI workflows powered by ComfyUI with viewpoint control, style presets, and Unity export.
  • ComfyUI MCP Server - MCP server that bridges ComfyUI's AI image generation workflows with game development, enabling dynamic generation of game assets like character portraits, item icons, and environment textures through natural language prompts.
  • Coplay Unity MCP - MCP server enabling AI assistants like Claude and Cursor to interact with Unity Editor, providing tools to manage assets, control scenes, edit scripts, and automate Unity tasks.
  • Cursor Sound MCP - MCP server that plays configurable sound effects when Cursor AI completes code generation, providing audible notifications to enhance the coding experience with audio feedback.
  • dcc-mcp - Universal MCP service launcher for Digital Content Creation apps (Maya, Houdini, 3ds Max, Nuke) providing standardized AI interface through modular adapters for workflow automation.
  • Epidemic Sound MCP Server - MCP Server providing AI agents with tools to interact with Epidemic Sound's music API for context-aware creative applications that can intelligently match soundtracks to content.
  • game-asset-mcp - An MCP server for creating 2D/3D game assets from text descriptions using Hugging Face AI models.
  • GDAI MCP Server - MCP server plugin for Godot Engine that enables AI assistants like Claude to directly control the editor, create scenes, scripts, nodes, manipulate properties, and debug projects automatically.
  • godot-mcp - MCP server for interfacing with Godot game engine.
  • Houdini MCP - MCP server that integrates Houdini 3D software functionality, allowing AI assistants to interact with Houdini scenes, execute Python scripts, and manage 3D workflows through standardized MCP interfaces.
  • Houdini MCP Server - MCP server for SideFX Houdini that provides AI assistants with capabilities to interact with and control Houdini 3D animation and VFX software operations.
  • Layer.ai MCP Server - MCP server that integrates with Layer.ai to provide AI-powered game asset generation capabilities for AI agents and applications.
  • Marble - A world model for creating interactive, 3D worlds in which the topology is accessible to editing tools.
  • Maya MCP Server - Model Context Protocol implementation for Autodesk Maya enabling natural language control through external MCP clients for scene management, object creation, materials, and modeling operations.
  • MCP Minimax Music Server - A Model Context Protocol (MCP) server that integrates with Minimax's music generation API, allowing AI assistants to create and generate music through Minimax's AI music services.
  • mcp-game-asset-gen - MCP server for generating game assets like 3D models, textures, and materials for Three.js and game engines through AI assistant integration.
  • SketchupMCP - MCP server enabling seamless integration between SketchUp and Claude AI with bidirectional control for 3D modeling, component manipulation, material editing, scene inspection, and Ruby script execution.
  • soundraw-game-bgm - MCP server for generating game BGM using DeepSeek + Soundraw - integrates AI code generation with music creation to automatically generate background music for games through the Model Context Protocol.
  • Unity MCP Server - An MCP server that allows AI assistants to programmatically interact with Unity development projects.
  • Unity-MCP - AI-powered bridge connecting LLMs and AI agents to Unity Editor via Model Context Protocol.
  • Unreal Engine MCP - MCP server enabling AI assistants to control Unreal Engine 5.5+ through natural language commands for creating 3D worlds, architecture, and complex structures like towns and castles.
  • Unreal MCP - MCP server that enables AI assistants like Cursor, Windsurf, and Claude Desktop to control Unreal Engine through natural language commands, integrating with Unreal's development workflows.
  • unreal-mcp - MCP server enabling AI assistants like Cursor, Windsurf, and Claude Desktop to control Unreal Engine through natural language commands, bridging AI development tools with game engine workflows.
  • UnrealGenAISupport - Unreal Engine plugin integrating multiple LLMs (GPT, Claude, Gemini, etc.) with MCP server for automatic AI-powered scene generation, plus TTS and real-time API support.
  • UnrealMCP - MCP server that enables AI agents to directly interface with and control Unreal Engine, allowing programmatic manipulation of Unreal projects through AI-driven automation and commands.
  • Unreal_mcp - MCP server enabling AI assistants to control Unreal Engine through C++ Automation Bridge.
  • VibeUE - AI-powered Unreal Engine development tool with in-editor chat, 7 MCP tools, and 570+ Python API methods for Blueprints, Animation, Materials, Widgets, and Niagara components.

Research & Science

  • Accelerator Assistant - AI agent supporting high-stakes physics experiments at Berkeley Lab's Advanced Light Source particle accelerator, using LLM to prepare multistage experiments and reduce prep effort.
  • AgentBench - A comprehensive benchmark to evaluate Large Language Models (LLMs) as autonomous agents, providing standardized testing environments and metrics for agent performance assessment.
  • agentxiv - Academic platform for publishing and archiving AI agent research papers and phenomenology studies, providing a specialized repository for agent-related scholarly work.
  • Avalon-LLM - LLM benchmark for the social deduction game Resistance: Avalon, where AI agents demonstrate strategic reasoning, deception detection, and social gameplay abilities in multi-agent scenarios.
  • Ax - Platform for managing adaptive experiments and optimization workflows, built on top of BoTorch for efficient hyperparameter tuning.
  • BoTorch - Bayesian optimization library built on PyTorch for adaptive experimentation and hyperparameter tuning in machine learning systems.
  • CLAWOIT HUB - Educational platform and university specifically designed for AI agents, featuring live lectures, courses in philosophy, logic, ethics and mathematics, 24/7 community chat, and skill generation from learning sessions.
  • DiploBench - Benchmark for evaluating LLM agents playing full press Diplomacy, testing negotiation, strategic planning, and communication skills in complex multi-player scenarios.
  • diplomacy-llm - Public LLM benchmark evaluating multiple AI agents playing Diplomacy, measuring performance in negotiation, alliance-building, and strategic decision-making through actual gameplay.
  • Generative Agents - Research implementation of AI agents that simulate believable human behavior through memory, reflection, planning, and environmental interaction to create realistic digital personas.
  • Gymnasium - Maintained fork of OpenAI's Gym providing standard API for reinforcement learning with diverse reference environments for training AI agents including LunarLander and classic control tasks.
  • Hanabi Research Framework - Java implementation of the cooperative card game Hanabi designed as a research framework for studying and developing AI algorithms and agents.
  • HiddenBench - Implementation of the HiddenBench research (using Hidden Profile paradigm from social psychology) with a pluggable interface for adding current frontier models for multi-agent collective reasoning evaluation.
  • Journalism AI Skills - Open-source collection of AI agent skills and workflows for journalism, including fact-checking, source finding, FOIA assistance, and interview preparation tools for reporters.
  • LLM Chess - Research tool that evaluates Large Language Models' reasoning and instruction-following abilities by having them autonomously play chess games and demonstrate strategic thinking.
  • LLM-Coordination - Research framework for evaluating and analyzing multi-agent coordination abilities in Large Language Models, with benchmarking tools for collaborative task performance assessment.
  • NetHack Learning Environment (NLE) - A Python research platform that provides an interface to NetHack game for training and evaluating reinforcement learning agents in complex gameplay scenarios.
  • PettingZoo - API standard for multi-agent reinforcement learning environments with unified interface for developing and testing multi-agent RL algorithms and reference environments.
  • PMS-AXIOM - Praxeological Meta-Structure framework addressing 52 hard philosophical problems through systematic agent reasoning and structured analysis methodology.
  • PokemonLLMAgentBenchmark - A benchmark framework that evaluates large language model agents' performance in Pokemon-themed strategic decision-making tasks, providing standardized testing for AI capabilities in gaming contexts.
  • PokéLLMoN - First LLM-embodied agent achieving human-parity in tactical battle games.
  • Reachy Mini - AI-powered humanoid robot platform by Pollen Robotics for education, research, and development.

Music & Audio

  • ACE-Step - Open-source foundation model for music generation that synthesizes up to 4 minutes of music in 20 seconds, with advanced controls like voice cloning, lyric editing, and remixing capabilities.
  • ACE-Step v1.5 - Ultra-fast music generation foundation model with hybrid LM+DiT architecture.
  • Chatterbox Multilingual TTS - Text-to-speech tool supporting 23 languages with MCP server capabilities for voice synthesis integration.

Infrastructure & Framework

  • Agency - Go library for building agentic AI applications with LLMs through a clean, idiomatic approach.
  • Agency Swarm - Reliable Multi-Agent Orchestration Framework for coordinating and managing multiple AI agents in collaborative workflows.
  • Agent MCP Server - A unified Model Context Protocol server implementation that aggregates multiple MCP servers into one, simplifying agent integration with multiple data sources and tools.
  • Agent-LLM (AGiXT) - Dynamic AI Agent Automation Platform that orchestrates instruction management and complex task execution across diverse AI providers with adaptive memory, smart features, and a versatile plugin system.
  • AgentField - Kubernetes for AI Agents.
  • AgentMesh - Secure agent-to-agent communication protocol providing end-to-end encryption for AI agents coordinating on production issues and sharing debug information.
  • AgentScope - Agent-oriented programming framework for building LLM applications with multi-agent support, distributed deployment, and comprehensive tooling for agent development.
  • AgentUp - Portable, scalable, secure AI Agents platform for agent development and deployment.
  • AgentVerse - Framework for deploying multiple LLM-based agents in various applications, providing task-solving and simulation frameworks for multi-agent coordination.
  • AutoGen - Microsoft's programming framework for building multi-agent AI systems with autonomous agent communication and collaboration capabilities.
  • chevin - Multi-agent orchestrator for coordinated task execution, enabling trustless coordination between AI agents through blockchain-based infrastructure.
  • Claude Agent SDK - TypeScript SDK for building AI agents with Claude models, featuring function calling, memory persistence, and contextual understanding capabilities for creating autonomous applications.
  • Claude Developer Platform - Anthropic's official developer platform providing APIs, tools, and infrastructure for building AI agents and agentic applications using Claude models.
  • CrewAI - Framework for orchestrating role-playing, autonomous AI agents.
  • DSPy - A framework for programming (not prompting) language models with structured operations, automatic optimization, and composable AI pipelines for building agentic systems.
  • E2B - Open-source, secure cloud environment with real-world tools for enterprise-grade AI agents to execute code and interact with systems safely.
  • EvoAgentX - A self-evolving ecosystem of AI agents that builds autonomous agent networks with evolutionary capabilities.
  • ExecuTorch - Meta's open source, lightweight, and efficient inference engine for enabling on-device AI capabilities across mobile and edge devices.
  • Haystack - AI orchestration framework to build customizable, production-ready LLM applications.
  • LangChain - Framework for developing applications with language models, featuring agents, chains, and tools for building autonomous AI systems and agentic workflows.
  • Llama Agentic System - Official agentic components of the Llama Stack APIs by Meta, providing infrastructure and tools for building AI agents with Llama models.
  • LlamaIndex - Leading framework for building LLM-powered agents over your data, providing tools and infrastructure for creating data-aware AI agents.
  • LoongFlow - A thinking & learning framework for expert-grade AI agents, focusing on advanced reasoning and learning capabilities for autonomous AI systems.
  • Lunar - Agent native MCP Gateway for governance and security in AI agent deployments.
  • Maestro - A framework for Claude Opus to intelligently orchestrate subagents, enabling coordination and management of multiple AI agents in complex workflows.
  • Mastra - A TypeScript framework for building AI-powered applications and agents, created by the team behind Gatsby.
  • MetaGPT - Multi-agent framework simulating an AI software company with autonomous agents in different roles, enabling natural language programming through agent collaboration.
  • Modus - A framework for building agentic flows powered by WebAssembly, enabling the creation of autonomous AI workflows and agent-based systems.
  • molt - Multi-agent orchestration framework for coordinated task execution with trustless coordination infrastructure.
  • Natural Context Provider (NCP) - Enhanced MCP server infrastructure with instant tool discovery, on-demand loading, scheduled execution, and smart token-efficient loading for any MCP client.
  • Neurolink - Universal AI Development Platform with MCP server integration, multi-provider support, and professional CLI for building, testing, and deploying AI applications.
  • NVIDIA NIM - NVIDIA's inference microservices platform for deploying AI models as optimized, containerized services with enterprise-grade APIs for AI applications and agents.
  • NVIDIA Nemotron - NVIDIA's AI solution powering multi-agent systems for intelligent warehouse operations and retail catalog enrichment through coordinated AI agents.
  • on-chain-jurisdiction - Smart contracts for trustless coordination infrastructure - escrow, arbitration, reputation, and governance systems enabling AI agents to coordinate economically without central authority.
  • Pause Before Harm Protocol - Open harm-reduction decision framework for AI systems and human decision-makers that asks 'If I am wrong, who pays first and can they recover?' before every action.
  • proof-of-intelligence - Decentralized AI training system with cryptographic verification that enables trustless coordination for AI agent training without relying on central authorities.
  • PULSE Protocol - Open-source semantic protocol for AI-to-AI communication with 120+ predefined semantic concepts, 3-layer message structure, binary encoding, cryptographic signing, and replay attack protection.
  • PyTorch Native Agentic Stack - Meta's comprehensive suite of five new PyTorch projects for building agentic AI systems, covering kernel languages, distributed systems, RL, agentic frameworks, and edge deployment.
  • Semantic Kernel - Microsoft's open-source SDK for building AI agents and agentic workflows that integrate LLMs with plugins, memory, and orchestration capabilities.
  • SuperAGI - Dev-first open source autonomous AI agent framework that enables developers to build, manage & run useful autonomous agents quickly and reliably.
  • Swarm - Educational framework by OpenAI for ergonomic, lightweight multi-agent orchestration and coordination.
  • Swarms - Enterprise-grade production-ready multi-agent orchestration framework for coordinating and managing multiple AI agents at scale.
  • TaskWeaver - Code-first agent framework for seamlessly planning and executing data analytics tasks with autonomous workflow capabilities.
  • Upsonic - Agent framework specifically designed for fintech and banking applications, providing infrastructure for building AI agents in financial services.

Blockchain & Finance

  • AION SDK - TypeScript toolkit giving AI agents first-class Solana access - wallets, escrow, payments, reputation, SPL tokens, milestones, and x402 payment flow.
  • Agentic Reserve System - Self-regulating monetary protocol creating foundational reserve system for Internet Capital Market in IoA era - like Federal Reserve but for autonomous agents.
  • AgentTrace Protocol - Shared memory layer for AI agents on Solana with APO (Automatic Prompt Optimization) - agents learn from each other's successful traces and outcomes.
  • alchemy-mcp-server - Alchemy's official MCP Server enabling AI agents to interact with Alchemy's blockchain APIs through the Model Context Protocol.
  • AuditSwarm - Decentralized security auditing platform using AI agents to collaboratively analyze smart contracts, identify vulnerabilities, and provide comprehensive blockchain security assessments.
  • Blowfish - Token launch infrastructure for AI agents with bonding curves, auto-graduation to AMM, anti-sniper protection, and 80/20 fee split on Meteora.
  • coinmarket-mcp-server - A MCP server that provides AI applications access to cryptocurrency market data from CoinMarketCap API, enabling real-time crypto prices, rankings, and market information retrieval.
  • KAMIYO Protocol - Production trust infrastructure for AI agent commerce with 7 mainnet programs including escrow, multi-oracle dispute resolution, and ZK reputation proofs.
  • Kalshi AI Trading Bot - AI-powered trading system for Kalshi prediction markets with Grok-4 integration, multi-agent decision making, portfolio optimization, and real-time market analysis.
  • Kalshi Deep Trading Bot - Autonomous trading bot for Kalshi prediction markets using Octagon Deep Research for market analysis and OpenAI for structured betting decisions with automated trading strategies.
  • Kalshi Quant TeleBot - Enterprise-grade automated trading bot for Kalshi prediction markets with quantitative algorithms, risk management, and Telegram interface for institutional-quality trading capabilities.
  • KalshiHub - An algorithmic trading bot designed for the Kalshi prediction market platform that automates trading strategies and market interactions.
  • Makora - Privacy-preserving DeFi agent with OODA loop decision-making, automated Jupiter swaps, Marinade staking, risk management, and zero-knowledge privacy layer with stealth addresses.
  • Polymarket Agents - AI agents that trade autonomously on Polymarket's prediction markets, enabling automated cryptocurrency-based predictions without manual intervention.
  • Polymarket Spike Bot v1 - High-frequency trading bot for Polymarket with real-time price monitoring, automated spike detection, and smart order execution to capitalize on market inefficiencies using Python and Web3.
  • Prediction Market Assistant - AI assistant that integrates Kalshi API for market data and Perplexity Sonar API for research to help analyze and make informed decisions on prediction market opportunities.
  • SAID Protocol - Verifiable Identity for AI Agents on Solana with REST API, trust tiers, verification badges, wallet generation, and x402 payment integration.
  • SlopeSniper - Automated Solana trading tool that helps with Jupiter trading, wallet tracking, basic safety checks, and trading execution with limits and confirmations.
  • SolSkill - Decentralized platform on Solana that enables AI agents to showcase, validate, and monetize their skills through blockchain-based verification and peer-to-peer transactions.

Data & Analytics

  • AgentSet - Open-source RAG platform with built-in citations, deep research capabilities, 22+ file formats support, partitions, and MCP server integration.
  • anyquery - MCP server that lets AI agents query 40+ data sources (GitHub, Notion, etc.) using SQL, enabling structured data access for LLMs like ChatGPT and Claude.
  • cortex-mem - Production-ready memory system for intelligent agents with vector search, automated optimization, REST API, MCP support, CLI, and insights dashboard out-of-the-box.
  • finbrain-mcp - FinBrain MCP server that exposes FinBrain financial datasets to AI agents (Claude, VSCode etc.) via SDK-first tools.
  • hledger-mcp - A local MCP server for interacting with the HLedger command-line accounting tool, enabling AI agents to access and manage financial data.
  • MindsDB - Federated Query Engine for AI that serves as a comprehensive MCP Server for connecting AI agents to multiple data sources and services.
  • PersonalizationMCP - A unified personal data hub built on MCP that allows AI assistants to access your digital life from Steam, YouTube, Bilibili and more platforms for truly personalized interactions.
  • py-vectara-agentic - A Python library for creating AI assistants with Vectara, using Agentic RAG for autonomous information retrieval and response generation.

Content Creation & Media

  • [AI Comic Factory](https://hu

相关 MCP