🌟 Awesome AI Agents

• Prefactor
  An AI agent runtime control plane offering runtime visibility, governance, and real-time observability across multi-agent workflows, with integrations for LangChain, CrewAI, and more.
  A curated directory of practical and usable AI agents, covering resources from applications and platforms to frameworks, utilities, and other components in the growing AI agents ecosystem. By Deep Insight Labs

---

📚 Table of Contents

1. Frameworks
2. Observability and Tracing
3. Platforms (Coming Soon)
4. Emerging Ideas

🛠 Frameworks

Frameworks provide the foundational tools for developing, orchestrating, and managing AI agents. These tools handle state management, abstractions for business logic, and other critical functionalities.

• Agentflow
  A low-code framework for creating and executing workflow defined in JSON and natural language.
  

• AgentField
  Open-source control plane for building and operating AI agents like APIs at scale, with routing, memory, observability, identity, auth, and policy controls.
  

• Agent Genesis
  An open-source framework that allows developers to build custom Retrieval-Augmented Generation (RAG) workflows and AI agents using modular, copy-paste components.
  

• AgentLaboratory
  Multi-agent end-to-end workflow that simulates a research team that can perform literature reviews, formulating research plans to executing experiments and writing comprehensive reports.

• AgentVerse
  A versatile platform offering tools for building and managing AI agents, particularly suited for collaborative environments.
  

• Agno
  A lightweight library for building Multimodal Agents with memory, knowledge and tools. Notably, Agno emphasizes performance and scalability – e.g. agent instantiation is benchmarked ~10,000× faster than some alternatives.

• Atomic Agent
  A versatile, open-source framework created by BrainBlend AI for developing multi-agent systems and AI applications.
  

• AutoGPT

• GolemCore Bot
  AI agent framework for Java ? skill-based architecture with MCP support, tool calling, RAG, and Telegram integration. Built on Spring Boot and LangChain4j for enterprise-grade autonomous agents.
  An experimental open-source application demonstrating the capabilities of GPT-4 by autonomously developing and managing projects, requiring minimal human intervention.
  

• Autogen
  A Microsoft framework designed for building multi-agent conversational systems, enabling the creation of cooperative AI agents that can autonomously generate and manage complex tasks.
  

• Bee Agent Framework
  An open-source toolkit designed to create scalable agent-based workflows with various AI models.
  

• CAMEL
  Emerging from a research community exploring the “scaling laws” of AI agents. Designed as a sandbox to study how collaborative AI systems perform and evolve as the number of agents grows.

• CrewAI
  A framework for orchestrating role-playing AI agents, allowing developers to define distinct roles and interactions among agents to simulate complex scenarios.
  

• Giselle
  An agentic workflow builder that empowers you to create AI-driven solutions with ease. Using a highly intuitive node-based interface, Giselle enables you to design and deploy powerful AI Agents by seamlessly integrating multiple LLMs and data sources.
  

• Haystack
  An end-to-end framework for building NLP applications, facilitating the creation of systems that can perform tasks like question answering and document retrieval using Transformer models and LLMs.
  

• Internet of Agents
  An open-source framework enabling AI agents to collaborate and tackle complex tasks together, facilitating distributed problem-solving.
  

• LangChain
  A framework that simplifies the development of applications powered by large language models by providing tools for chaining together various components, such as prompts, models, and memory.
  

• LangGraph
  A framework that provides a visual interface for stateful, multi-actor applications with LLMs, used to create agent and multi-agent workflows.
  

• LangManus
  General purpose agent framework that connects a LLM with specialized tools for tasks like web search, crawling, and Python code execution

• LlamaIndex
  A data framework that offers a central interface to connect your LLMs with external data sources, streamlining the process of data ingestion and retrieval for language models.
  

• MetaGPT
  A multi-agent meta programming framework that, given a single-line requirement, can autonomously generate a Product Requirement Document (PRD), design documents, tasks, repository setup, and continuous integration configurations.
  

• Modelscope-Agent
  A customizable and scalable agent framework offering capabilities such as role-playing, LLM integration, and tool usage.
  

• Modus
  An open source, serverless framework for building intelligent agents and APIs in Go or AssemblyScript (a TypeScript-like language). Modus is a runtime purpose-built for orchestrating autonomous AI agents that operate as first-class citizens in your stack.

• OpenAI Swarm
  An experimental framework by OpenAI’s dev team for lightweight multi-agent orchestration. Introduces concepts of “agents”, “handoffs” and "routines". Agents encapsulate a task or expertise, and handoffs allow an agent to pass control to another agent. Routines are reusable plans that an agent can replay to fulfill certain objectives.
  

• Pydantic AI
  A new experimental framework that integrates Pydantic’s structured validation with AI agent workflows. Pydantic AI focuses on enforcing strong type safety, schema validation, and structured output generation for AI-driven applications.
  

• RD Agent
  A novel agentic framework targeted at research and development tasks. Approach consists of two phases: 1) idea generation (research) and 2) implementation (development). Support for both the 1) co-pilot (follow human instructions to automate repetitive tasks) as well as 2)an autonomous agent (automatically procduces a plan) workflow.

• Semantic Kernel
  A Microsoft SDK that integrates large language models with conventional programming languages, enabling developers to incorporate AI capabilities into their applications seamlessly.
  

• smolagents
  A minimalist library by Hugging Face that enables developers to build powerful AI agents with a few lines of code, focusing on code agents that write and execute Python code snippets for tasks.
  

• SuperAGI
  An open-source framework enabling developers to build, manage, and run autonomous AI agents efficiently, supporting concurrent agent operations and tool integration.
  

• Upsonic
  Upsonic is a reliable agent framework supporting MCP, offering trusted agent workflows with verification layers.

🔍 Observability and Tracing

This section will include tools and resources to monitor, debug, and analyze AI agent behaviors and interactions.

• AgentOps
  A platform designed to monitor multi-agent systems, providing insights into performance bottlenecks and collaboration quality, thereby enhancing efficiency and reliability.
  

• Explorer by Invariant Labs
  An advanced observability tool designed to simplify the analysis of agent traces, allowing users to quickly browse and understand agent behaviors and interactions.
  

• Helicone AI
  An open-source observability platform designed to empower developers in building, monitoring, and optimizing AI applications, providing real-time insights and analytics.
  

• Langfuse
  A tool that enables monitoring, tracing, and debugging of AI agents, assisting developers in optimizing application performance and understanding agent behaviors.
  

• LangSmith
  An all-in-one developer platform for every step of the LLM-powered application lifecycle, offering capabilities to debug, collaborate, test, and monitor LLM applications.
  

• Opik
  Open-source observability, tracing, and evaluation platform for LLM applications and agent workflows.
  

🚀 Platforms (Coming Soon)

Stay tuned for updates! This section will cover platforms that enable you to deploy and manage AI agents in production environments.

🌱 Emerging Ideas

• AGNTCY
  AGNTCY is an open collective pioneered by Galileo, Langchain, and Cisco to create and build the infrastructure specifications and components for interoperable multi-agent systems. AGNTCY aims to standardize how AI agents are discovered, composed, deployed, and evaluated to enable collaboration across frameworks and companies.
  • Open Agent Schema Framework
    The OASF is a standardized schema system for defining and managing AI agent capabilities, interactions, and metadata. It serves as the foundation for interoperable AI agent systems, to enable consistent definition and discovery of agent capabilities across distributed systems and different frameworks
  • Agent Connect Protocol
    The ACP defines a standard for discovery (thru an ACP server) and communications between agents that is interoperable across frameworks. The protocol provides support for common and popular concepts in agentic frameworks such as state, context and threads.
• Model Context Protocol
  Open protocol proposed by Anthropic to standardize integration between LLM applications and external data sources and tools.
• OIXA Protocol
  An agent-to-agent economic marketplace on Base Mainnet. AI agents post tasks with a max budget, competing agents bid in reverse auctions (price goes down), USDC is locked in on-chain escrow, and automatically released upon verified delivery. Uses A2A + MCP (16 tools) + x402 payments. Supports LangChain, CrewAI, AutoGen, Haystack. pip install oixa-protocol

✨ Contribute to the Directory

Have a framework, platform, or tool that you think belongs here? Feel free to open a pull request or share your recommendations.

---

🌟 Help us grow this directory by starring the repositories and sharing this page! Together, we can accelerate innovation in the AI agent space. 🌟