Awesome Agents is a curated list of open-source tools and products on the theme of building AI agents. The list is maintained by kyrolabs on GitHub and is intended to provide developers with an index of various resources needed to build AI agents.
In terms of purpose, the list is aimed at developers and teams who want to build, test, evaluate, and deploy AI agents. It categorizes tools by functional area, making it easy for users to find the appropriate framework or platform according to specific needs.
Its main capability is reflected in its categorization structure. The list includes the following categories: Frameworks, Testing and Evaluation, Software Development, Conversational / General Agents, Game / Simulation, Memory - Knowledge Management, and Automation Browser. Each category includes multiple open-source projects with brief descriptions.
In the frameworks category, the list includes many projects such as LangChain, LlamaIndex, Haystack, Semantic Kernel, AutoGen, AG2, CrewAI, MetaGPT, Lagent, AgentScope, Swarm, Mastra, VoltAgent, smolagents, Pydantic Deep Agents, and others. These frameworks cover multiple languages such as Python, TypeScript, Go, and C#, and support capabilities such as multi-agent orchestration, tool calling, memory management, and MCP integration.
The testing and evaluation category includes tools such as Voice Lab, Open-RAG-Eval, EvoAgentX, Arize-Phoenix, Manifest, agent-qa, Latitude, and ClawMetry, used for agent testing, evaluation, observability, and cost tracking.
The software development category includes projects such as MetaGPT, OpenHands, GPT Pilot, Aider, Devika, RepoAgent, DSPy, SWE Agent, Cline, OpenCode, and Stakpak, aimed at coding agents.
In addition, the list also includes tools in categories such as conversational/general agents, games/simulation, memory and knowledge management, and automation browsers. At the top of the list, a new list called Awesome ADE is also mentioned, focusing on open-source agentic development environments, used for running coding agents in parallel, TUI and tmux tools, desktop and web workspaces, worktree and container isolation, and more.
In terms of usage, users can directly visit the GitHub repository to browse the categorized directory and click on the links of each project to view details. The list itself does not provide installation or run commands; for specific usage, refer to each project's own documentation.
Applicable scenarios include: developers selecting AI agent frameworks, researchers looking for evaluation tools, teams looking for coding agents or automation tools, and understanding the current state of the AI agent ecosystem.
In terms of limitations, the list is a resource index and does not guarantee the quality, maintenance status, or security of the included projects. Some projects may have stopped being maintained or may have changed. The list content is updated over time, and information such as specific features, pricing, and platform support should be based on each project's official materials.
