AutoSwarm

Overview

AutoSwarm turns fragmented agents into one coordinated team. Given a task once, it selects agents by capability, shares context and memory, connects complementary tools and workflows, bridges data and application ecosystems, runs tasks in parallel, coordinates collaboration, and verifies and synthesizes results into one coherent answer.

Key Features

• Selects the right agents based on their capabilities.
• Shares context and memory across agents.
• Connects complementary tools and workflows and bridges data and application ecosystems.
• Runs tasks in parallel and coordinates agent collaboration.
• Verifies and synthesizes results into one coherent answer.
• A lightweight plug-in lets any supported agent serve as the entry point.

Use Cases

• Summarizing and comparing large document sets by delegating to multiple agents.
• Coordinating coding, research, browsing, office work, writing, computer use, and social networking tasks across agents.
• Running multi-agent tasks in parallel for faster execution.

Getting Started

• Give AutoSwarm to an agent you already use and ask it to install or update AutoSwarm for that agent, reuse existing configuration, and verify which locally installed agents are available.
• After installation, start a task by asking the agent to use AutoSwarm, for example to read 100 documents and summarize key approaches and disagreements.
• For a first installation, obtain the official repository, install the package with uv tool install --force . or python -m pip install ., and run autoswarm setup --host with the appropriate host ID.

Deployment & Requirements

• Requires Python 3.10+ or uv for installation; setup installs the Skill, runtime, and host integration.
• Works across Windows, macOS, and Linux.
• Supports local command execution.

Before You Adopt

• License: MIT. Review its terms before using, modifying, or distributing the project.
• AutoSwarm does not provide an automatic updater; updates are handled manually.
• Users should not reuse the archived v1 environment.
• Probing providers spends model calls to check a reply and native resume.
• Version checks alone only establish executable discovery, not full provider verification.
• Installing a missing provider CLI requires asking first and configuring only the agents needed.