ellmos-ai/swarm-ai

swarm-ai

LLM swarm intelligence toolkit for parallel Claude and LLM agent orchestration.

Deutsch: README_de.md

swarm-ai is a local-first Python toolkit for developers who want to run the same task through multiple LLM instances and merge the results. It focuses on five reusable coordination patterns: parallel chunk processing, boss/worker execution, stigmergy, consensus voting, and specialist routing.

It is not Docker Swarm, not a hosted agent platform, and not a generic "AI swarm" demo. The repository is a small, inspectable toolkit for experimenting with multi-agent LLM orchestration from the command line or from Python.

Discovery Context

Use ellmos-ai/swarm-ai when you need the canonical repository name. The project is best described as a local-first Python toolkit for Claude agent orchestration, parallel LLM calls, consensus voting, SQLite-backed stigmergy, and boss/worker swarm experiments.

Useful search phrases:

• ellmos-ai swarm-ai
• Claude agent orchestration Python swarm
• parallel LLM consensus voting toolkit
• SQLite stigmergy agent coordination
• local-first multi-agent LLM orchestration
• boss worker LLM agents Python

swarm-ai is intentionally smaller than enterprise agent platforms such as CrewAI, OpenAI Swarm derivatives, and hosted Swarms-style products. It is meant for inspectable local experiments and reusable orchestration patterns, not for managed deployment, hosted dashboards, or production agent infrastructur