Powered by Claude

Multi-Agent AI Workflows,
Made Simple

Spin up a mesh of Claude agents in seconds. Coordinate complex tasks, analyze repositories, and automate intelligent pipelines — all from your terminal.

npx claude-flow@alpha hive init --topology mesh --agents 5

Everything you need to orchestrate AI agents

Claude Flow gives you a production-ready multi-agent platform in one command.

⬡

Mesh Topology

Agents communicate peer-to-peer in a resilient mesh network. No single point of failure.

🔄

Dynamic Scaling

Spin up any number of agents on demand and distribute workloads automatically.

🧠

Claude-Powered

Every agent is backed by Anthropic's Claude API — fast, safe, and highly capable.

📦

Zero Config

One npx command and you're running. No servers to provision, no infra to manage.

🔍

Repo Analysis

Point a worker swarm at any GitHub organisation to analyse, diagnose, and summarise repos at scale.

🔒

Secure by Default

API keys stay in your environment. No credentials leave your machine.

How it works

  1. 01

    Initialise the hive

    Run npx claude-flow@alpha hive init to bootstrap your agent cluster with the topology and agent count of your choice.

  2. 02

    Write your worker

    Use the claude_worker.py template to define exactly what each agent should do — from repo analysis to code generation.

  3. 03

    Run and collect results

    Execute your worker script. Agents fan out across the workload in parallel and return structured insights.

Quickstart

Up and running in under two minutes.

1. Bootstrap the hive
npx claude-flow@alpha hive init --topology mesh --agents 5
2. Set your API key
export ANTHROPIC_API_KEY="sk-ant-..."
3. Run the worker
python claude_worker.py

See the AI reason in real time

Watch Gemma 3 think through your request step-by-step, emit structured tool calls, and deliver a final answer — all in your browser, no setup required.

Open the chat →