ruvnet/agentic-flow

Easily switch between alternative low-cost AI models in Claude Code/Agent SDK. For those comfortable using Claude agents and commands, it lets you take what you've created and deploy fully hosted agents for real business purposes. Use Claude Code to get the agent working, then deploy it in your favorite cloud.

What it solves

Agentic-Flow is a production-ready orchestration framework for AI agents. It addresses the complexity of managing large swarms of specialized agents, reducing the overhead of routing tasks to the correct agent, lowering LLM operational costs, and improving the accuracy and speed of multi-agent consensus and memory retrieval.

How it works

The system utilizes a "Self-Optimizing Neural Architecture" (SONA) for sub-millisecond pattern learning and adaptive routing. It integrates AgentDB for advanced memory management, employing various attention mechanisms (Flash, Linear, Hyperbolic, and MoE) and Graph Neural Networks (GNN) to refine queries and improve recall. The framework provides 66 pre-built specialized agents (e.g., coders, reviewers, planners) and 213 MCP tools for coordination and integration with platforms like GitHub.

Who it’s for

  • Developers looking to build complex multi-agent systems with pre-built specialized agents and high-performance local code editing.
  • Businesses seeking to reduce LLM token costs and infrastructure overhead while scaling agentic workflows.
  • Researchers interested in implementing and testing state-of-the-art attention mechanisms and graph-aware reasoning in agent swarms.

Highlights

  • SONA Adaptive Learning: Sub-millisecond pattern retrieval and LoRA fine-tuning for real-time adaptation.
  • Advanced Attention: Support for Flash Attention (up to 7.47x speedup), Linear Attention for long sequences, and Hyperbolic Attention for hierarchical swarms.
  • Lage Agent Library: 66 self-learning agents that improve autonomously via a "ReasoningBank" memory system.
  • GNN Query Refinement: Improves information recall by 12.4% through graph-context integration.
  • Cost Efficiency: Intelligent model routing (e.g., between Sonnet and Haiku) to reduce costs by up to 60%.
  • Extensive Tooling: 213 MCP tools for swarm initialization, memory management, and GitHub integration.

Related

  • Project
    kyegomez/swarms

    A production-ready multi-agent orchestration framework that provides pre-built architectures for coordinating multiple AI agents to solve complex tasks.

  • Project
    nullpointexception-i/agent-sphere

    AgentSphere is a Java‑Spring/React platform that runs LLM‑driven ReAct loops, supports multi‑agent sub‑runs, real browser automation via a Chrome extension, model routing, persistent multi‑level memory, human‑in‑the‑loop clarification, OIDC SSO, and an embeddable chat widget.

  • Project
    agentscope-ai/agentscope

    A production-ready agent framework that provides an SDK and a FastAPI-based service to build, deploy and scale multi-agent applications with built-in RAG and tool management.

  • Project
    catlog22/maestro-flow

    An intent-driven multi-agent orchestration framework that automates the full software development lifecycle by dynamically planning and executing complex command chains across multiple LLM backends.

  • Dispatch
    Intelligence is Free, Now What? Data Systems for, of, and by Agents

    BAIR researchers propose a new framework for data systems redesigned for agentic workloads, agentic state management, and agent-driven system synthesis as AI inference costs approach zero.