Frameworks
General-purpose agent and LLM application frameworks (the app layer, not harnesses per se).
| # | Project | Stars | Tier | OSS | Description |
|---|---|---|---|---|---|
| 1 | n8n workflowlocaltypescript | 205k | complex | ⚠️ Fair-code | Fair-code workflow engine with 400+ nodes and native AI nodes; the self-hosted Zapier that actually does agents and LangChain. |
| 2 | AutoGPT memoryevalspython | 187k | complex | ⚠️ Polyform-SU | The original autonomous loop: goal in, agent iterates with tools and memory; Forge is the dev framework, Benchmark the eval harness. |
| 3 | Dify low-coderagpython | 157k | complex | ⚠️ Fair-code | One-stop LLM app platform: visual workflows, RAG pipeline, 50+ tools, model management; "ship from prototype to prod" in a single UI. |
| 4 | langflow low-codepython | 155k | complex | ✅ | Low-code UI to build and deploy LangChain/LangGraph flows; visual DAG editor and one-click run. |
| 5 | langchain python | 147k | complex | ✅ | Chains, tools, retrievers, and agents; the usual entry point for "add tools to an LLM" in Python/JS. |
| 6 | browser-use browserpython | 116k | slightly complex | ✅ | Python web-agent **harness**: natural-language goals become browser actions, driven directly over the Chrome DevTools Protocol (it dropped Playwright in August 2025). The biggest community in the browser-agent category. |
| 7 | Flowise low-codetypescript | 55.5k | complex | ⚠️ Apache+CLA | Drag-and-drop LangChain UI; deploy flows without code. The low-code sibling to Langflow, with a different component and hosting story. |
| 8 | llama-index ragpython | 52.2k | complex | ✅ | Data-centric: indexing, RAG, and query engines; agent abstractions sit on top of your data pipelines. |
| 9 | agno memoryevalspython | 42.3k | complex | ✅ | Python agents with memory, knowledge bases, tools, and structured outputs; continues the PhiData-era product line under the Agno name—production apps, evals, and pipelines. |
| 10 | langgraph workflowpython | 42k | slightly complex | ✅ | State-machine graphs over LLM steps; checkpointing, human-in-the-loop, and durable execution so workflows survive restarts. |
| 11 | semantic-kernel python | 28.6k | complex | ✅ | Microsoft's plugin and planner layer for LLMs; C#, Python, Java; strong on enterprise auth and orchestration. |
| 12 | mastra typedtypescript | 28.2k | slightly complex | ⚠️ Elastic-2.0 | TypeScript-first; agents, tools, and workflows with a single runtime and minimal boilerplate. |
| 13 | Haystack memoryragpython | 26.6k | complex | ✅ | Open-source orchestration framework for context-engineered LLM apps: modular pipelines and agent workflows with explicit control over retrieval, routing, memory, and generation—closer to LangChain's territory than a coding-agent harness. |
| 14 | letta memorypython | 24.8k | mostly simple | ✅ | Python agent runtime with tool use and control flow; lean API; stateful agents with long-horizon memory. |
| 15 | Stagehand browsertypescript | 24.6k | slightly complex | ✅ | Browserbase's SDK for browser agents: natural-language actions (act, extract, observe) and deterministic Playwright code mix in one script, so agent flexibility and repeatable automation live in the same **harness**. |
| 16 | Google ADK evalssandboxpython | 21.6k | complex | ✅ | Google's official Agent Development Kit: code-first Python toolkit for building, evaluating, and deploying agents. Optimized for Gemini but model-agnostic; deploys to Cloud Run / Vertex AI; ships a dev UI with eval and a code-execution sandbox. |
| 17 | rasa voicepython | 21.3k | complex | ✅ | Conversational AI stack (NLU, dialogue, actions); long-standing OSS choice for chat and voice bots. |
| 18 | botpress low-codetypescript | 14.9k | complex | ✅ | Visual bot builder and runtime; multi-channel, open-source alternative to commercial bot platforms. |
| 19 | R2R visionragworkflowpython | 8k | complex | ✅ | RAG-first: hybrid search, knowledge graphs, multimodal; the framework for "production RAG" when you care more about retrieval than chat UI. |
| 20 | agent-squad multi-agent | 7.8k | slightly complex | ✅ | AWS-originated orchestrator (now under 2FastLabs): intent classification, streaming, SupervisorAgent; "agent-as-tools" so one agent delegates to a squad. |
| 21 | AgentVerse multi-agentpython | 5.1k | complex | ✅ | Task-solving and simulation envs for multi-LLM agents; deploy many agents in custom environments without building infra from scratch. |
| 22 | youtu-agent | 4.6k | mostly simple | ❓ | Tencent Cloud's agent framework: a minimal tool-calling **harness** designed to perform well with open-source models, positioned as a lighter alternative to heavier orchestration frameworks. |
| 23 | Bee Agent Framework mcpmulti-agentpythontypescript | 3.4k | complex | ✅ | Python + TypeScript, LF AI–backed; MCP/ACP, workflows, Requirement Agent; the one that pushes "production multi-agent" without LangChain. |
| 24 | AgentStack | 2.2k | slightly complex | ✅ | Scaffolds full agent projects; plugs in CrewAI, LangGraph, OpenAI Swarm, LlamaStack and wires AgentOps observability from day one. |
| 25 | AgentSilex python | 456 | super simple | ✅ | ~300 lines of readable agent code on top of LiteLLM; the "I want to see the whole loop" option for learning or minimal production. |
| 26 | SuperAgentX multi-agentpython | 203 | mostly simple | ✅ | Lightweight multi-agent orchestrator with an AGI-angle; minimal surface, docs-first, for teams that want orchestration without the kitchen sink. |