Curated, ranked list of AI agent harnesses (the runtimes that close the loop between a stateless model and the outside world), orchestration frameworks, and harness techniques for reliable agentic systems.
160+ harnesses · 12 categories · weekly-rescored · MCP-ready :
claude mcp add agent-harnesses -- uvx agent-harnesses-mcp| Project | Stars | Category | Tier | OSS |
|---|---|---|---|---|
| OpenClaw typescriptmulti-agent | 392k | Personal agent runtimes | complex | ✅ |
| superpowers memorycliide | 302k | Coding harness configs and SDKs | complex | ✅ |
| Hermes memorypythonprovider-agnostic | 255k | Personal agent runtimes | slightly complex | ✅ |
| opencode mcpprovider-agnosticclituitypescript | 216k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| n8n workflowlocaltypescript | 208k | Frameworks | complex | ⚠️ |
| AutoGPT memoryevalspython | 188k | Frameworks | complex | ⚠️ |
| Anthropic Skills | 182k | Coding harness configs and SDKs | mostly simple | ⚠️ |
| Dify low-coderagpython | 159k | Frameworks | complex | ⚠️ |
| langflow low-codepython | 156k | Frameworks | complex | ✅ |
| langchain python | 148k | Frameworks | complex | ✅ |
| GStack typescript | 137k | Coding harness configs and SDKs | slightly complex | ✅ |
| Codex sandboxprovider-agnosticcli | 130k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| browser-use browserpython | 118k | Frameworks | slightly complex | ✅ |
| pi provider-agnostictuirust | 116k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| addyosmani/agent-skills workflowide | 108k | Coding harness configs and SDKs | mostly simple | ✅ |
| Gemini CLI mcpclitypescript | 107k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| claude-mem memory | 102k | Memory and state | slightly complex | ✅ |
| MCP Servers mcpmemorytypescript | 91.3k | Plugins, MCPs, CLI tools | mostly simple | ✅ |
| OpenHands memorybrowsersandboxpython | 91.2k | Coding agent products (IDEs, CLIs, full suites) | complex | ⚠️ |
| DeerFlow memorymulti-agentsandboxpython | 84.1k | Research and task-specific harnesses | complex | ✅ |
| Headroom mcprag | 75.8k | Progressive disclosure harnesses | mostly simple | ✅ |
| MetaGPT multi-agentpython | 70.9k | Multi-agent and orchestration | complex | ✅ |
| Cline idetypescript | 70.5k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| Open Interpreter clipython | 68.6k | Coding agent products (IDEs, CLIs, full suites) | mostly simple | ✅ |
| Mem0 memorypython | 67.4k | Memory and state | slightly complex | ✅ |
| AnythingLLM ragtypescript | 67.3k | Personal agent runtimes | complex | ✅ |
| Context7 mcptrainingtypescript | 63.2k | Plugins, MCPs, CLI tools | super simple | ✅ |
| LiteLLM provider-agnosticpython | 61.5k | Libraries and SDKs | mostly simple | ✅ |
| autogen multi-agentpython | 61.5k | Multi-agent and orchestration | complex | ✅ |
| crewAI python | 59.8k | Multi-agent and orchestration | complex | ✅ |
| OpenManus multi-agentpython | 58.7k | Multi-agent and orchestration | complex | ✅ |
| awesome-claude-code | 55.9k | Coding harness configs and SDKs | super simple | ⚠️ |
| Flowise low-codetypescript | 55.5k | Frameworks | complex | ⚠️ |
| goose mcprust | 55.4k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| chrome-devtools-mcp mcpbrowsertypescript | 53.8k | Plugins, MCPs, CLI tools | mostly simple | ✅ |
| llama-index ragpython | 52.6k | Frameworks | complex | ✅ |
| aider mcpclipython | 49.6k | Plugins, MCPs, CLI tools | slightly complex | ✅ |
| nanobot mcpmemorylocalpython | 49.1k | Personal agent runtimes | mostly simple | ✅ |
| CowAgent memorypython | 47.5k | Personal agent runtimes | slightly complex | ✅ |
| langgraph workflowpython | 43.4k | Frameworks | slightly complex | ✅ |
| agno memoryevalspython | 42.8k | Frameworks | complex | ✅ |
| wshobson/agents multi-agentcliide | 41k | Coding harness configs and SDKs | super simple | ✅ |
| awesome-cursorrules ide | 41k | Progressive disclosure harnesses | super simple | ✅ |
| Playwright MCP mcpvisionbrowsertypescript | 38.3k | Plugins, MCPs, CLI tools | mostly simple | ✅ |
| Khoj python | 37.7k | Personal agent runtimes | complex | ✅ |
| continue idetypescript | 36.2k | Plugins, MCPs, CLI tools | complex | ✅ |
| Langfuse evalstypescript | 36k | Observability and eval-ops | slightly complex | ✅ |
| DeepSeek-Reasonix memoryclituitypescript | 35.8k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| oh-my-pi browserprovider-agnosticcliiderust | 35.8k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| ChatDev python | 34.5k | Multi-agent and orchestration | slightly complex | ✅ |
| github-mcp-server mcp | 33.7k | Plugins, MCPs, CLI tools | slightly complex | ✅ |
| cognee memoryragworkflowpython | 32.5k | Memory and state | slightly complex | ✅ |
| Graphiti (Zep) memoryragworkflowpython | 32k | Memory and state | slightly complex | ✅ |
| context-mode mcpmemorysandbox | 30.7k | Progressive disclosure harnesses | mostly simple | ⚠️ |
| Composio sandboxtool-discoverypythontypescript | 30.6k | Libraries and SDKs | complex | ✅ |
| deepagents multi-agentsandboxpythontypescript | 30.4k | Libraries and SDKs | slightly complex | ✅ |
| gpt-researcher multi-agentpython | 30.2k | Research and task-specific harnesses | complex | ✅ |
| openai-agents-python python | 30.1k | Multi-agent and orchestration | mostly simple | ✅ |
| smolagents sandboxpython | 29.9k | Libraries and SDKs | mostly simple | ✅ |
| mastra typedtypescript | 28.9k | Frameworks | slightly complex | ⚠️ |
| crush memoryclitui | 28.7k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| semantic-kernel python | 28.7k | Frameworks | complex | ✅ |
| qwen-code sandboxclitypescript | 28.6k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| MLflow evalspython | 28.4k | Observability and eval-ops | complex | ✅ |
| vibe-kanban | 28.4k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| beads memory | 28k | Memory and state | mostly simple | ✅ |
| Symphony sandbox | 27.7k | Coding agent products (IDEs, CLIs, full suites) | complex | ✅ |
| Kilo Code mcpcliidetypescript | 27.6k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| planning-with-files memory | 27.5k | Coding harness configs and SDKs | mostly simple | ✅ |
| vercel/ai provider-agnostictypescript | 27.4k | Libraries and SDKs | slightly complex | ✅ |
| Haystack memoryragpython | 26.8k | Frameworks | complex | ✅ |
| Stagehand browsertypescript | 25.7k | Frameworks | slightly complex | ✅ |
| letta memorypython | 25.2k | Frameworks | mostly simple | ✅ |
| agents.md idetypescript | 25k | Progressive disclosure harnesses | super simple | ✅ |
| MCP Python SDK mcppython | 24.6k | Plugins, MCPs, CLI tools | mostly simple | ✅ |
| Roo Code mcpworkflowidetypescript | 24.3k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| Opik evalspython | 22.7k | Observability and eval-ops | slightly complex | ✅ |
| Prime Agent memorymulti-agentclitypescript | 22.1k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| Google ADK evalssandboxpython | 21.9k | Frameworks | complex | ✅ |
| rasa voicepython | 21.4k | Frameworks | complex | ✅ |
| jcode mcpmemoryprovider-agnosticclirust | 21k | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| pydantic-ai mcptypedprovider-agnosticpython | 20.7k | Libraries and SDKs | slightly complex | ✅ |
| SWE-agent memoryevalspython | 20.6k | Coding harness configs and SDKs | slightly complex | ✅ |
| Eliza memorymulti-agenttypescript | 19.6k | Personal agent runtimes | complex | ✅ |
| Agent Zero memorymulti-agentbrowsersandboxpython | 19.5k | Personal agent runtimes | slightly complex | ✅ |
| Agent Lightning evalstrainingpython | 18.7k | Evaluation and benchmarking harnesses | complex | ✅ |
| OpenHarness (HKUDS) memorymulti-agent | 16k | Personal agent runtimes | complex | ✅ |
| eigent multi-agentlocal | 15.6k | Coding agent products (IDEs, CLIs, full suites) | complex | ✅ |
| QM memorysandboxtypescript | 15.4k | Personal agent runtimes | complex | ✅ |
| cc-haha memorymulti-agenttypescript | 15.1k | Coding agent products (IDEs, CLIs, full suites) | complex | ✅ |
| botpress low-codetypescript | 15k | Frameworks | complex | ✅ |
| AutoResearchClaw multi-agent | 14.7k | Research and task-specific harnesses | complex | ✅ |
| E2B sandboxpython | 14.4k | Libraries and SDKs | slightly complex | ✅ |
| Microsoft Agent Framework multi-agentworkflowpython | 14.2k | Multi-agent and orchestration | slightly complex | ✅ |
| MCP TypeScript SDK mcptypescript | 13.6k | Plugins, MCPs, CLI tools | mostly simple | ✅ |
| Arize Phoenix evalspython | 11.9k | Observability and eval-ops | slightly complex | ⚠️ |
| OpenJarvis memorylocalpython | 11.5k | Personal agent runtimes | slightly complex | ✅ |
| hive multi-agentpython | 11.1k | Multi-agent and orchestration | complex | ✅ |
| MCP Inspector mcptypescript | 11.1k | Plugins, MCPs, CLI tools | super simple | ✅ |
| omnigent sandboxidepython | 10.9k | Multi-agent and orchestration | complex | ✅ |
| get-shit-done clipython | 10.6k | Coding harness configs and SDKs | mostly simple | ✅ |
| PraisonAI multi-agentpython | 9.3k | Multi-agent and orchestration | mostly simple | ✅ |
| strands-agents mcpmulti-agenttypedpython | 8.9k | Libraries and SDKs | mostly simple | ✅ |
| MiroThinker evals | 8.4k | Research and task-specific harnesses | slightly complex | ✅ |
| Claude Agent SDK mcpmemorypythontypescript | 8.3k | Coding harness configs and SDKs | complex | ✅ |
| R2R visionragworkflowpython | 8k | Frameworks | complex | ✅ |
| Steel memorybrowserlocal | 7.8k | Libraries and SDKs | slightly complex | ✅ |
| agent-squad multi-agent | 7.8k | Frameworks | slightly complex | ✅ |
| MCP Registry mcp | 7.4k | Plugins, MCPs, CLI tools | slightly complex | ✅ |
| Agent Governance Toolkit sandboxpython | 6.4k | Plugins, MCPs, CLI tools | slightly complex | ✅ |
| agents-cli evalscli | 6.1k | Coding harness configs and SDKs | mostly simple | ✅ |
| SWE-bench evalssandboxpython | 6k | Evaluation and benchmarking harnesses | slightly complex | ✅ |
| Cloudflare Agents memorytypescript | 5.8k | Libraries and SDKs | slightly complex | ✅ |
| skillhub localcliide | 5.2k | Coding harness configs and SDKs | mostly simple | ✅ |
| AgentVerse multi-agentpython | 5.2k | Frameworks | complex | ✅ |
| AG2 multi-agentpython | 5k | Multi-agent and orchestration | complex | ✅ |
| mcp-context-forge mcppython | 4.6k | Plugins, MCPs, CLI tools | complex | ✅ |
| youtu-agent | 4.6k | Frameworks | mostly simple | ✅ |
| Agent Sandbox memorysandboxlocal | 4.3k | Libraries and SDKs | slightly complex | ✅ |
| openai-agents-js multi-agentvoicetypescript | 3.9k | Libraries and SDKs | slightly complex | ✅ |
| AgentBench evalssandboxragworkflowpython | 3.8k | Evaluation and benchmarking harnesses | complex | ✅ |
| Bee Agent Framework mcpmulti-agentpythontypescript | 3.4k | Frameworks | complex | ✅ |
| inspect_ai evalssandboxpython | 3k | Evaluation and benchmarking harnesses | complex | ✅ |
| cocoindex-code mcpcli | 2.9k | Plugins, MCPs, CLI tools | mostly simple | ✅ |
| agent-vault | 2.4k | Plugins, MCPs, CLI tools | mostly simple | ⚠️ |
| AgentStack | 2.2k | Frameworks | slightly complex | ✅ |
| Meta-Harness | 1.7k | Coding harness configs and SDKs | slightly complex | ✅ |
| WebArena python | 1.6k | Evaluation and benchmarking harnesses | complex | ✅ |
| Docker MCP Gateway mcpsandboxcli | 1.6k | Plugins, MCPs, CLI tools | slightly complex | ✅ |
| AIlice sandboxpython | 1.4k | Personal agent runtimes | slightly complex | ✅ |
| WebVoyager evalsvision | 1.1k | Evaluation and benchmarking harnesses | slightly complex | ✅ |
| ClawBench evalsvisionsandboxpython | 1.1k | Evaluation and benchmarking harnesses | complex | ✅ |
| Terminal-Bench evalsclipython | 925 | Evaluation and benchmarking harnesses | slightly complex | ✅ |
| agent-qa mcpmemorysandboxclitypescript | 910 | Evaluation and benchmarking harnesses | slightly complex | ⚠️ |
| swe-smith trainingpython | 804 | Evaluation and benchmarking harnesses | slightly complex | ✅ |
| SWE-Gym evalstrainingpython | 749 | Evaluation and benchmarking harnesses | slightly complex | ✅ |
| ARC-AGI-2 | 745 | Evaluation and benchmarking harnesses | super simple | ✅ |
| inspect_evals evalssandbox | 706 | Evaluation and benchmarking harnesses | slightly complex | ✅ |
| open-harness mcpmulti-agenttypescript | 619 | Libraries and SDKs | slightly complex | ✅ |
| langgraph-bigtool tool-discoverypython | 557 | Progressive disclosure harnesses | slightly complex | ✅ |
| RepoMaster workflowpython | 555 | Coding harness configs and SDKs | slightly complex | ❓ |
| claw-code-agent mcprustpythontypescript | 549 | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ❓ |
| AgentBox sandboxtypescript | 538 | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| Proliferate multi-agentsandboxidetypescript | 516 | Coding agent products (IDEs, CLIs, full suites) | complex | ✅ |
| MCP-Zero tool-discovery | 515 | Progressive disclosure harnesses | complex | ✅ |
| AgentSilex python | 458 | Frameworks | super simple | ✅ |
| openagents | 457 | Research and task-specific harnesses | complex | ✅ |
| AutoHarness memorymulti-agentprovider-agnosticpython | 452 | Coding harness configs and SDKs | super simple | ✅ |
| arc-agi-benchmarking evalsprovider-agnosticpython | 365 | Evaluation and benchmarking harnesses | mostly simple | ✅ |
| AgentRL trainingpython | 360 | Multi-agent and orchestration | complex | ✅ |
| SuperAgentX multi-agentpython | 203 | Frameworks | mostly simple | ✅ |
| ToolGen tool-discoverypython | 184 | Progressive disclosure harnesses | complex | ✅ |
| VitaBench | 182 | Evaluation and benchmarking harnesses | complex | ✅ |
| LoopTroop typescript | 163 | Coding harness configs and SDKs | mostly simple | ✅ |
| AgencyBench evalssandboxpython | 102 | Evaluation and benchmarking harnesses | complex | ✅ |
| letta-evals memorypython | 83 | Evaluation and benchmarking harnesses | mostly simple | ✅ |
| Talon mcpmemoryclitypescript | 79 | Personal agent runtimes | slightly complex | ✅ |
| YYLO multi-agenttypescript | 64 | Coding agent products (IDEs, CLIs, full suites) | slightly complex | ✅ |
| SUPER sandboxpython | 58 | Evaluation and benchmarking harnesses | slightly complex | ✅ |
| ToolRAG mcptool-discovery | 34 | Progressive disclosure harnesses | mostly simple | ✅ |
| puppeteer-real-browser-mcp mcpbrowsertypescript | 26 | Plugins, MCPs, CLI tools | mostly simple | ✅ |
| TRAIL | 24 | Evaluation and benchmarking harnesses | mostly simple | ✅ |
| Community-curated agent lists | 16 | Libraries and SDKs | super simple | ✅ |
| pmstack evals | 9 | Coding harness configs and SDKs | super simple | ✅ |
| Better-OpenCodeMCP mcptypescript | 9 | Plugins, MCPs, CLI tools | mostly simple | ✅ |
| agentlog memoryclipython | 1 | Plugins, MCPs, CLI tools | super simple | ✅ |