AutoGen
Provider: Microsoft Research
Open-source programming framework for building AI agents and facilitating cooperation among multiple agents to solve tasks.
Free
Use AgentOpen-Source Framework AI Agents in AI Agent Development
7 Open-Source Framework AI agent(s) in AI Agent Development.
BabyAGI
Provider: Yohei Nakajima
Experimental framework for a self-building autonomous agent, leveraging LLMs for task management and ideation.
Free
Use AgentDify
Provider: LangGenius
Dify is an open-source platform (also available as managed cloud/SaaS) for building LLM-powered agents and RAG workflows. It offers a visual canvas to design “apps” that combine LLM reasoning, retrieval from knowledge docs, prompts, tool integrations (Zapier via MCP), and deployment—all with built-in logs and observability. It supports OpenAI, Anthropic, Llama2, Hugging Face, and customizable agents.
Free from $59/month
Use AgentHaystack
Provider: deepset GmbH
Haystack is an open-source Python framework for building production-ready, agent-like applications. It enables developers to construct pipelines and agents that use chat-based LLMs and external tools (e.g., search APIs, data connectors) in iterative loops with state management, tool selection, and exit conditions. Components include retrievers, readers, generators, routers, tool-invokers, and evaluation modules.
Free deepset Cloud
Use AgentKimi K2
Provider: Moonshot AI
A 1T-parameter Mixture-of-Experts model with 32B active parameters at inference. Trained with model-scaling innovations (“MuonClip”) and optimized for reasoning and tool use—benchmarked strongly on SWE‑bench, coding, math, and reasoning tasks.
Free
Use AgentLangroid
Provider: Langroid
Python framework to compose multi‑agent systems with tools, memory, and robust control flows.
Open‑source
Use AgentVoltAgent
Provider: VoltAgent
VoltAgent is an open‑source TypeScript framework (MIT‑licensed) for building, orchestrating, and monitoring AI agents. It provides core modules (agents, memory, tool integrations), multi‑agent orchestration, and a visual console (VoltOps) for tracking execution steps, tool calls, and agent reasoning. It supports LLM‑agnostic providers (OpenAI, Anthropic, Google, etc.) and integrates with observability tools like Langfuse.
Free
Use Agent