AIAny
AI Agent2025
Icon for item

Claude-Flow

Wraps Claude Code as an MCP server that orchestrates 100+ specialized agents into self-organizing swarms — hierarchical, mesh, or adaptive consensus — backed by persistent vector memory, coordination hooks, and secure cross-machine federation.

Introduction

The bottleneck for agentic coding stopped being the model a while ago — it's coordination. A lone Claude Code session forgets across runs, can't split work cleanly, and has no safe way to hand tasks between machines. Claude-Flow's bet is that the missing piece is infrastructure, not a cleverer prompt: it bolts a memory layer, a scheduler, and a federation protocol onto Claude Code (and Codex) so agents act like a team instead of a chat loop.

What Sets It Apart
  • Runs as an MCP server and native Claude Code plugin, so you keep the host you already use rather than adopting yet another agent framework.
  • 100+ role-specialized agents self-organize into swarms with hierarchical, mesh, or adaptive-consensus topologies — you describe the goal, not the task graph.
  • Persistent, HNSW-indexed vector memory carries context across sessions, so the next run isn't a cold start and prior decisions stay reachable.
  • Federation coordinates agents across separate machines under zero-trust security — the part that matters once one laptop stops being enough.
Who It's For

Great fit if you already live in Claude Code or Codex and keep hitting the ceiling of single-session, single-machine work: multi-step refactors, paired test-and-security passes, or projects that run for days. Look elsewhere if you want a one-shot coding helper — the swarm, memory, and federation machinery is overhead you won't touch. It's also a fast-moving, solo-led project (powered by Cognitum.One) rather than a vendor-backed platform, so expect churn across its frequent releases.

Information

  • Websitegithub.com
  • OrganizationsCognitum.One
  • Authorsruvnet
  • Published date2025/06/02

More Items

Enables RL post-training with million-token prompts under a fixed GPU budget by evaluating shared prompt state without autograd, retaining only minimal model state, and replaying short response branches; instantiated as GRPO and demonstrated on Qwen3.6-27B and GLM-5.2 up to multi-million token execution.

GitHub
AI Client2026

Transforms unstructured documents into strongly-typed Knowledge Abstracts with one CLI command, extracting entities and relations into graphs, hypergraphs, and spatio‑temporal structures. Includes 80+ templates, multiple RAG engines, local vLLM support, Obsidian export and an MCP server.

GitHub
AI Infra2026

Defines OpenTelemetry semantic conventions for generative AI telemetry — spans, metrics, and events for GenAI clients, the Model Context Protocol (MCP), and provider-specific integrations. Includes YAML models, human-readable docs, and reference implementations to standardize observability across GenAI deployments.