Multi-agent tools and frameworks, compared
Multi-Agent Guide lists 105 multi-agent tools and frameworks. Each one has a page with the same structure: what it is, key facts, features, architecture, MCP / A2A / AG-UI support, what it fits, a quickstart, pros and cons, and alternatives. GitHub numbers are refreshed daily and every fact links to its source.
Multi-Agent Tool Selection Guide — $9 one-time
Full dataset (CSV / JSON / XLSX / Notion), comparison matrix and a PDF selection guide, updated monthly for 12 months. Every page on this site stays free.
Start here
- All tools
One filterable table: type, language, license, pricing, protocols.
- Protocol support matrix
MCP, A2A and AG-UI per tool, with evidence links and check dates.
- Agent teams in Claude Code: what they are, how they differ from subagents, when to use them
Topic page.
- New tools and releases
Recent launches and major versions, each with an official source.
Best for
- Best multi-agent tools for coding agents
- Best multi-agent frameworks for customer support
- Best multi-agent frameworks for enterprise governance
- Best multi-agent frameworks for Java and the JVM
- Best tools for running multiple coding agents in parallel on one repository
- Best multi-agent frameworks for research agents
- Best multi-agent tools for self-hosted and local use
- Best multi-agent frameworks for TypeScript
Comparisons
- Agent2Agent (A2A) Protocol vs Model Context Protocol (MCP)
- AG2 vs AutoGen
- Claude Code agent teams vs Claude Code subagents
- AgentScope vs AutoGen
- CrewAI vs Agno
- Cursor vs Devin
- DeerFlow vs OpenHands
- Google ADK (Python) vs LangGraph
- Haystack vs LlamaIndex
- Herdr vs cmux
- LangGraph vs CrewAI
- Letta vs LangGraph
- Mastra vs VoltAgent
- MetaGPT vs ChatDev
- Microsoft Agent Framework vs AutoGen
- Multica vs Symphony
- OpenAI Agents API vs LangGraph
- OpenAI Agents SDK (Python) vs LangGraph
- OpenClaw vs Hermes Agent
- Orca vs Conductor
- Orca vs Superset
- Paseo vs T3 Code
- Pydantic AI vs OpenAI Agents SDK (Python)
- Ruflo vs Claude Code subagents
Tools by GitHub stars
| Tool | Type | Languages | License | Pricing | Pattern | MCP | A2A | AG-UI | Stars |
|---|---|---|---|---|---|---|---|---|---|
| OpenClaw | Harness | TypeScript | unknown | Open source, free | Supervisor | Yes evidence | Yes evidence | Unknown | 390,998 |
| Superpowers | Orchestrator | Markdown, Shell, JavaScript, TypeScript | MIT | Open source, free | Supervisor | Unknown | Unknown | Unknown | 293,495 |
| ECC | Orchestrator | JavaScript, TypeScript, Shell, Python | MIT | Open core | Other | Yes evidence | Unknown | Unknown | 270,233 |
| Hermes Agent | Harness | Python | MIT | Open core | Supervisor | Yes evidence | Yes evidence | Unknown | 250,364 |
| DeepSeek Harness | Harness | TypeScript, Python | MIT | Open source, free | Supervisor | Yes evidence | Unknown | Unknown | 241,122 |
| OpenCode | Harness | TypeScript | MIT | Open core | Supervisor | Yes evidence | Unknown | Unknown | 211,180 |
| n8n | Platform | TypeScript, JavaScript, Python | unknown | Open core | Supervisor | Yes evidence | Unknown | Unknown | 206,389 |
| AutoGPT | Platform | Python, TypeScript | unknown | Open core | Supervisor | Yes evidence | Unknown | Unknown | 187,631 |
| Dify | Platform | Python, TypeScript, JavaScript | unknown | Open core | Graph | Yes evidence | Unknown | Unknown | 157,623 |
| agency-agents | Orchestrator | Markdown, Shell | MIT | Open source, free | Crew / roles | Partial evidence | Unknown | Unknown | 155,538 |
| Langflow | Platform | Python, TypeScript | MIT | Open source, free | Supervisor | Yes evidence | Yes evidence | Yes evidence | 155,406 |
| Claude Code | Harness | TypeScript, Python | unknown | Hosted, paid | Supervisor | Yes evidence | Unknown | Unknown | 148,737 |
| LangChain | Framework | Python | MIT | Open core | Supervisor | Yes evidence | Yes evidence | Partial evidence | 147,331 |
| gstack | Orchestrator | TypeScript, Markdown, Shell | MIT | Open source, free | Pipeline | Partial evidence | Unknown | Unknown | 134,616 |
| Codex CLI | Harness | Rust, TypeScript, Python | Apache-2.0 | Open core | Supervisor | Yes evidence | Unknown | Unknown | 127,433 |
New
- Paseo 0.10.2: fixes for OpenCode v2 sessions
- AgentScope v2.0.9 adds an experimental SOP module and TeamPipeline
- CrewAI 1.15.23: evaluation of the last traced run and tracing changes
- Multica v0.6.0: issue deliverables, attachment viewer and issue wakeup v2
- Google ADK (Python) v2.10.0: experimental skill lifecycle, MongoDB toolset, eval efficiency metrics