Ruflo is an agent orchestration platform designed to enhance Claude AI with multi-agent swarms, self-learning memory, and secure collaboration. It provides tools and plugins for building sophisticated conversational AI systems.
Ruflo operates using a self-learning/self-optimizing agent architecture. The system involves user interaction via CLI/MCP, routing, swarms, agents, memory, and LLM providers in a learning loop. This enables agents to self-organize, learn from tasks, and remember across sessions. The core is powered by WASM kernels written in Rust.
Ruflo boasts a rich set of features including self-learning, federated communications, and enterprise security. It provides a plugin marketplace with options for core orchestration (e.g., ruflo-swarm, ruflo-autopilot), memory and knowledge (ruflo-agentdb, ruflo-rag-memory), intelligence and learning (ruflo-intelligence, ruflo-ruvllm), and more.
Agent Federation creates secure workspaces for agents across different machines or organizations. Agents can discover each other, prove their identities, and collaborate on tasks with PII automatically stripped. This is facilitated using mTLS, ed25519, and behavioral trust scoring mechanisms.
Ruflo offers a web UI (beta) at flo.ruv.io which provides a multi-model AI chat with Model Context Protocol (MCP) tool calling. The Goal Planner UI at goal.ruv.io allows users to set plain-English goals that Ruflo then decomposes into executable agent plans. The platform is self-hostable and also offers hosted demos for easy access.
Ruflo is a powerful platform for orchestrating AI agents with Claude, providing enhanced collaboration, security, and learning capabilities. Its modular plugin architecture and user-friendly interfaces, including a web UI and goal planner, make it a versatile tool for building advanced AI systems and autonomous workflows. While still under development, Ruflo offers significant potential for enterprises looking to leverage AI agent collaboration effectively.
Ruflo is an agent orchestration platform for Claude AI that enables the deployment of intelligent multi-agent swarms. It coordinates autonomous workflows and builds conversational AI systems with features like self-learning memory and enterprise-grade security.
Ruflo enhances agent collaboration through features like Agent Federation, which allows agents on different machines to discover, authenticate, and securely exchange work. PII is stripped before data leaves any node, ensuring data safety and compliance.
The Ruflo plugin marketplace includes plugins like ruflo-core for foundational server functions, ruflo-swarm for coordinating multiple agents, ruflo-agentdb for fast vector database management, and ruflo-federation for secure agent collaboration across different machines.
The Goal Planner UI is a front-end interface that allows users to input plain-English goals, which Ruflo then decomposes into executable agent plans. It utilizes a GOAP A* planner to find the shortest viable path and adaptively replans when the state changes.
You can try Ruflo by using hosted demos like the web UI at flo.ruv.io and the Goal Planner at goal.ruv.io. Alternatively, the platform is self-hostable via Docker, allowing for deployment on your own infrastructure.
Source: https://github.com/ruvnet/ruflo