ruvnet/agentic-flow
Production AI agent orchestration library with 66 self-learning agents, 213 MCP tools, and autonomous multi-agent swarm coordination.

Velocity · 7d
+1.2
★ / day
Trend
→steady
star history
Agentic-Flow is a TypeScript framework for orchestrating AI agents across multiple providers. It enables switching between Claude, Gemini, and OpenRouter models while managing a swarm of specialized agents that self-learn and optimize routing over time. The system includes an MCP server for Claude Code integration, background workers triggered by keywords, and a neural architecture component (SONA) for adaptive learning.