ruvnet/ruflo
ruflo turns a solo AI coding assistant into a team of coordinating agents
ruflo (formerly Claude Flow) is an execution layer that sits on top of Claude Code or Codex and lets multiple AI agents work together instead of one agent working alone. A single command, npx ruflo init, adds over 100 specialized agents, memory that persists across sessions, self-learning, and secure communication between agents on different machines. It's an open-source TypeScript repository with 68644 stars.
What it does
- The framing is 'Agent = Model + Harness': the AI model does the writing, while ruflo is the harness that gives it tools, memory, execution loops, sandboxes and security controls
- There are two install paths: a lightweight Claude Code plugin with a handful of slash commands, or the full npx ruflo init CLI setup with 98 agents, 60+ commands, an MCP server, and hooks
- Multiple agents can be organized into 'swarms' using different coordination structures (hierarchical, mesh, adaptive) with consensus mechanisms to work as a team
- Its vector memory system, AgentDB with an HNSW search index, measured roughly 1.9x faster search at 20,000 records and roughly 3.2x-4.7x faster at 5,000 records compared to brute-force search, with a recall@10 accuracy around 0.99
- A 'federation' feature lets agents on different machines or organizations discover each other, verify identity cryptographically, strip out personal data before sending messages, and collaborate securely
Why it matters
For development teams trying to automate repetitive coding, review, and security work, this shows a shift from a single AI assistant to a coordinated team of agents that remembers past work and improves over time. The reported 68644 GitHub stars and over 8.1 million ecosystem downloads indicate the scale of real-world adoption.
Terms in this repo
- harness · the execution framework that gives an AI model tools, memory, and control loops so it can actually perform tasks
- MCP (Model Context Protocol) · a standardized protocol that lets AI models communicate with external tools and servers
- swarm · a group of AI agents coordinated together like a team, each handling a different role
- HNSW · a search index technique for quickly finding similar items among large sets of vector data
- federation · a system where agents on separate machines or organizations verify each other's identity and share work securely
Repository description (English)
🌊 The original agent meta-harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
Open on GitHubTrending repos
- openai/codexOpenAI's coding agent, now runnable straight from your terminal
- cordiverse/cordisA TypeScript framework for snapping feature pieces on and off whenever you need them
- ripienaar/free-for-devA single repo lists every free-tier cloud, API and collaboration tool developers actually need
- Wei-Shaw/sub2apiA relay server that lets one Claude/OpenAI/Gemini/Grok subscription be shared by many users
- multica-ai/andrej-karpathy-skillsA single instruction file that stops AI coding assistants from making silent bad guesses
- eneskirca/nodetermA workspace that puts scattered terminal tabs and AI coding agents on one draggable map
- affaan-m/ECCA toolkit that teaches AI coding agents to work like disciplined engineers
- n8n-io/n8nn8n lets you build automation workflows and AI agents by dragging nodes on a canvas, with code as a fallback