AgentIndex icon
AgentIndex
ToolsMCP ServersHOTCategoriesTrendingNewBest ListsCompareGuides
Submit Tool
ToolsMCP ServersHOTCategoriesTrendingNewBest ListsCompareGuides
Home/
Multi-Agent/
ruflo
ruflo logo

ruflo

Active·★ 79.8k·MIT·Updated 2026-10-27
★ Most Popular

Ruflo (formerly Claude Flow) is an agent "meta-harness" for Claude Code and Codex. Its premise is that an agent is a model plus a harness, and Ruflo is the harness layer: it adds specialised agents, commands, skills, hooks and an MCP server so that agents can organise into swarms, remember across sessions and learn from earlier runs, while you keep working in Claude Code as usual.

It is aimed at heavy users of Claude Code or Codex who want more structure around long or parallel work: coordinating several agents on one task, tracking cost and tokens, and reviewing what each agent did. A visual console shows workflows, agents, models and cost, lets you replay and compare runs, and can be driven by Claude itself within permission levels you set (read, write, manage or full, with ask or auto approval). Federation lets agents on different machines communicate.

Getting started: in Claude Code 2.1.287 or later, add the `ruvnet/ruflo` plugin marketplace and install `ruflo-core` (plus the console and mods if wanted). For the full setup with hooks and the MCP server, run `npx ruflo@latest init wizard` in your project; other MCP clients can launch `npx -y ruflo@latest mcp start`.

This write-up is based on the project's README and public repository data. We have not hand-tested the tool. Data as of 2026-10-24.

License
MIT
Latest release
v3.61.1 · 2026-10-24
Last commit
2026-10-24
Contributors
80
Forks
9,281
Open issues and PRs
975
First published
2025-06-02
npm package
@claude-flow/codex 3.0.5 · 5,840 downloads last week

Source: GitHub, npm and PyPI public APIs.

#AI Orchestration#Multi-Agent System#Self-Learning AI#Claude Code Integration#Vector Database#Consensus Algorithms#WebAssembly (WASM)#Distributed Systems
$ Install
$ npx ruflo@latest init wizard
↗ Visit site★ GitHub
01

Features

01Adds a large set of specialised agents, commands, skills, hooks and an MCP server around Claude Code and Codex
02Agent swarms that self-organise and coordinate in the background through hooks
03Memory across sessions with local vector retrieval and learning from past runs
04Visual console to track workflows, agents, models, tokens and cost, with replay of runs
05Federation for agents on different machines to communicate
02

Why choose it

+Deep integration with Claude Code through plugins, hooks and a console, plus MCP support for Codex and other clients.
+Multi-agent coordination and cross-session memory without changing how you use Claude Code day to day.
+Console with cost and token tracking, run replay and adjustable control levels for the agent.
+Very actively developed and MIT licensed.
03

Trade-offs

!The full CLI install adds many files and hooks to your workspace (`.claude/`, `.claude-flow/`, CLAUDE.md and more); the plugin path is lighter but more limited.
!Mods run with your account's permissions and are not sandboxed; the project advises reading their code before installing.
!A very large surface (hundreds of MCP tools, many commands and agents) means a steep learning curve and more context overhead.
!Closely tied to Claude Code; hook support varies in other clients.
05

Maintenance check

ActiveChecked: 2026-10-25
Last commit
2026-10-25
Commits, last 90 days
100+
Releases, last 90 days
30
Latest release
v3.63.0 (2026-10-25)
Open issues
985
Recent issues with at least one comment
2/9
Commits per week, last 12 weeks
Commits per week, last 12 weeks: 63, 23, 3, 10, 55, 59, 126, 375, 630, 233, 214, 8

Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.

Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.

Most discussed open issues

  • #3189 [verification] HIGH: Witness signature verification fails — all platform manifests report every entry missing84 comments · 2026-09-05
  • #108 Claude Flow v2.0.0 - Comprehensive Functionality Review & Enhancement Proposals57 comments · 2025-07-04
  • #2156 [Dream Cycle 2026-05-27] intelligence: SR²AM 8B=120–355B via simulative planning — 95% token gap + capabilities,memory scan54 comments · 2026-05-27
  • #2970 [verification] HIGH: witness verify.mjs reports every manifest entry missing in source-only checkout — build artifacts unverifiable36 comments · 2026-08-11
  • #2036 track: ADR-121 — embeddings upgrade to ruvector 0.2.x + @ruvector family33 comments · 2026-05-17

Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.

04

Who it fits

Pick it if
  • You are a heavy Claude Code or Codex user and want several agents coordinated on one task, with memory across sessions.
  • You want a console that tracks cost and tokens and lets you replay and compare runs.
  • You want control levels (read, write, manage, full) with ask or auto approval.
Look elsewhere if
  • You want a light setup: the full CLI install adds many files and hooks to your workspace.
  • You cannot read mods before installing: they run with your permissions and are not sandboxed.
  • Context overhead matters: hundreds of MCP tools and many commands mean a steep learning curve.
06

How it differs

ralph-claude-code logovs ralph-claude-code★ 9.7k

Ralph keeps a single Claude Code loop running safely. Ruflo is a full harness for multi-agent swarms.

Full comparison →
07

Compatibility

Claude Code
Native
Verified via docs
Node.js
Node.js
Verified via docs
08

Quick start

1
$ npx ruflo@latest init wizard
09

Use cases

↳Running parallel Claude agent swarms for large-scale document processing
↳Building autonomous workflows that require inter-agent coordination and consensus
↳Deploying production AI pipelines with persistent vector memory across sessions
10

Alternatives

⚡Top VS:
VSralph-claude-code★9.7k→VSMassGen★1.1k→VSMetaGPT★71.0k→VSBettaFish★42.5k→
ralph-claude-code logo
ralph-claude-code★ 9.7k
Autonomous AI development loop for Claude Code with intelligent exit detection
VSCompare→
MassGen logo
MassGen★ 1.1k
🚀 MassGen is an open-source multi-agent scaling system that runs in your terminal, autonomously orchestrating frontier models and agents to collaborate, reason, and produce high-quality results. | Join us on Discord: discord.massgen.ai
VSCompare→
MetaGPT logo
MetaGPT★ 71.0k
🌟 The Multi-Agent Framework: First AI Software Company, Towards Natural Language Programming
VSCompare→
BettaFish logo
BettaFish★ 42.5k
Multi-agent opinion analysis assistant. Breaks information bubbles and predicts trends.
VSCompare→
omnigent logo
omnigent★ 11.1k
Omnigent is an open-source AI agent framework and meta-harness: orchestrate Claude Code, Codex, Cursor, Pi, and custom agents — swap harnesses without rewriting, enforce policies and sandboxing, and collaborate in real time from any device.
VSCompare→
claude-vibe-squad logo
claude-vibe-squad★ 164
Multi-model AI orchestration where behaviour is Markdown, not code. One coordinator routes scoped task packets to 71 role-based specialists across 5 model families (Codex / Claude / Gemini / Grok / Kimi) — each working in an isolated git worktree, each reviewed by a rival family before it lands. One tmux session, no server.
VSCompare→
musubi logo
musubi★ 11
Musubi (結び) — Ai Agent shared memory and thought layer. The braiding of threads between presences.
VSCompare→
neo-agent-brain logo
neo-agent-brain★ 12
Point a cross-model AI engineering team at your own codebases. The Agent OS keeps named AI maintainers' identity, memory and review across labs alive between sessions — the Brain Neo.mjs runs on itself. Multi-tenant: onboarding a codebase is a config entry, not a fork.
VSCompare→
See all alternatives →

Featured in these lists

  • Open-Source Multi-Agent Frameworks: CrewAI, Pydantic AI, Google ADK, Microsoft Agent Framework & More
  • Best Claude Code Workflow Tools, Skills and Harnesses

Comments

Log in to leave a comment

No comments yet. Be the first!

On this page
01Features02Why choose it03Trade-offs05Maintenance check04Who it fits06How it differs07Compatibility08Quick start09Use cases10Alternatives
ruvnet /
ruflo
↗

Agent meta-harness for Claude Code and Codex swarms

79.8k9.3kMIT
Last commit1d ago
StatusActive
CategoryMulti-Agent
Contributors80
Latest releasev3.61.1
Links
Documentation↗Discussion↗Issues↗Releases↗
Cloud Infrastructure
Ad

Deploy on DigitalOcean

Host your MCP servers, Docker containers, and AI agents with $200 free credit over 60 days.

Claim $200 Free Credit
© 2026 AgentIndex.app|Built by a 10-year iOS Developer.
Best by use caseAboutContactAdvertisePrivacy PolicyTerms of UseQYSGitHubBuy me a coffee ☕

Browse by Category

MCP ServersCode AssistantWorkflow AutomationRAG / Knowledge BaseMulti-AgentBrowser AutomationLLM InfraDev ToolingObservability

Not affiliated with Anthropic, OpenAI or Microsoft.