AutoSwarm README

The project's README from github.com/HKUDS/AutoSwarm. As one markdown file: README.md.

AutoSwarm logo AutoSwarm: Unify All Your Agents. Break Context Barriers.

简体中文

Paper Feishu WeChat

Download AutoSwarm for Windows Download AutoSwarm for Mac (Apple silicon) Download AutoSwarm for Mac (Intel)

New: AutoSwarm Desktop Released — All Your Agents in One Private Workspace

One window to coordinate agents, share context, review approvals, and track progress.

Watch all your agents work together for you. What AutoSwarm does

HKUDS%2FAutoSwarm | Trendshift

AutoSwarm Desk

Why AutoSwarm?

No single agent can do everything. You don’t need another agent—you need all your agents working as one.

Different agents come with different:

Today, you are forced to switch interfaces, repeat information, coordinate subtasks, and manually combine results.

AutoSwarm turns fragmented agents into one coordinated team.

Give AutoSwarm a task once. It unifies all your agents into one coordinated team:

🤖 Agents AutoSwarm Can Call


Codex


Claude Code


OpenCode


Cursor


GitHub Copilot CLI


Meta Muse Code


CodeBuddy


Hermes


Nanobot


DeepSeek Harness


Grok Build

Kimi Code logo
Kimi Code

AutoSwarm works across Windows, macOS, and Linux. Use AutoSwarm through:

The main host needs Skill support and local command execution. Collaborators need compatible local CLIs; your assistant can help connect missing ones.

🖥️ AutoSwarm Desk App

AutoSwarm Desk app

🚀 Easy to use

One Lightweight Plug-in. Any Agent as Your Entry Point. All Your Agent at Work.

How to Use AutoSwarm

Give AutoSwarm to any agent you already use and ask:

Install or update AutoSwarm for this agent. Reuse my existing configuration and verify which locally installed agents are available.

Your assistant will report which collaborators are ready. To add another, tell it the agent's name; it will handle setup and verification.

Then start any task with:

Use AutoSwarm to read these 100 documents and summarize the key approaches and disagreements.

AutoSwarm selects, connects, and coordinates the right agents for every task.

🤖 AutoSwarm in Action

https://github.com/user-attachments/assets/89a077a9-168e-441f-8aac-48de321aa968

📰 News

🗺️ Roadmap

-🔌 Expand Integrations — Support more agent products, models, tools, and native capabilities.
-🤝 Improve Coordination — Strengthen context sharing, communication, task routing, and parallel execution.
-🛡️ Enhance Reliability — Improve setup, monitoring, error recovery, verification, and overall user experience.

🌟 Citation

If you find this work helpful to your research, please kindly consider citing our paper.

@misc{li2026autoswarm,
      title={AutoSwarm: Automating Agent Swarms},
      author={Zhonghang Li and Zongwei Li and Chao Huang},
      year={2026},
      eprint={2610.33048}
}

For agents: How to use AutoSwarm

Installation and updates — instructions for the configuring agent (click to expand)

Install

  1. Inspect the current host and existing Python 3.10+ / CLI installations. Reuse native models, credentials, proxy settings and project context. Ask before installing a missing provider CLI; configure only the agents needed.

  2. For a first installation, use git clone to obtain the official repository, then install the package and Skill for the current host. Reuse an existing checkout when available. Example for Codex:

    git clone https://github.com/HKUDS/AutoSwarm.git
    cd AutoSwarm
    uv tool install --force .
    autoswarm setup --host codex
    

    Other host IDs: claude-code, opencode, cursor, copilot, muse, codebuddy, hermes, nanobot, dsh, grok, kimi. Without uv, use a dedicated Python 3.10+ environment and python -m pip install .; keep its interpreter available. Do not reuse the archived v1 environment.

  3. Read the installed SKILL.md at the path returned by setup. It contains the launcher argv and usage instructions. Use that exact launcher when the App cannot find autoswarm. Run agents in the actual host environment; if needed, merge an existing CLI's absolute argv into providers.<provider>.command in the reported AutoSwarm home/config.json. Preserve unrelated settings.

  4. Verify requested providers in the actual project, replacing the example provider and directory as appropriate:

    autoswarm doctor --agent codex --probe --cwd /path/to/project
    

    Repeat --agent for several providers. This spends model calls to check a reply and native resume; verified: true confirms the probe. Version checks only establish executable discovery. An empty native authentication list does not prove that login is required. Handle pending native approvals/questions under the installed Skill's rules, then finish verification.

Setup installs the Skill, its runtime and the host's integration. Follow the returned reload/trust instructions, then briefly name the verified collaborators and explain that the user can request others by name. If none are ready, propose connecting a suitable agent; obtain any missing installation authorization and handle configuration and verification. See configuration and recovery for native profiles, permissions and connection diagnostics.

Update an existing installation

Obtain the requested revision from the original source. For an existing Git checkout, preserve local edits, collect active delegated work, then run on the intended branch:

git pull --ff-only
uv tool install --reinstall .
autoswarm setup --host codex

Use the actual host, existing AutoSwarm home and original installation destinations. Put --home <existing-home> before setup when needed; retain custom --skill-dir, --settings-file and --plugin-dir values. Refresh each host installation in use. For pip installations, use the existing dedicated interpreter to run python -m pip install --force-reinstall ., then python -m autoswarm setup --host codex with the same options.

Run setup from the updated package: the installed Skill has its own runtime copy, so updating the checkout or package alone is insufficient. The old Skill launcher cannot fetch a new version. Keep configuration, records and native authentication, reload the host as required, read the refreshed Skill and verify the needed providers. Finish with the same brief collaborator summary as installation. There is no automatic updater. See upgrade details.

Use and recover

Follow the installed Skill to select agents, delegate ready work, collect results with wait --any, handle intermediate questions/approvals and verify deliverables before answering. Ordinary tasks use the current conversation; native wake-up is optional and needs a supported host binding.

Read only the relevant references: capabilities and data scope, communication, native interactions, model selection. Diagnose retained calls and logs before retrying uncertain operations.

Architecture · Contributing and validation · MIT License

AutoSwarm (HKUDS/AutoSwarm)

Searching Database…

What is this?AutoSwarm is a coordinator that lets the agents you already use (Codex, Claude Code, Cursor and others) work as one team. You give it a task once, and it picks the right agents, shares context between them and combines their results into one answer.
Is it a new AI agent or a layer on top of existing ones?It is not a new agent. It is a plug-in or Skill that you add to an agent you already use, and that host then connects your other agents as collaborators.
How do I get it running?Clone the official repository with git clone, install the package with python -m pip install . (or uv), then run python -m autoswarm setup --host codex. Reload the host as the instructions say, then ask the assistant which collaborators are ready.
What do I need first?Python 3.10 or newer, and a host agent that supports Skills and local command execution. Every other agent you want to use needs its own compatible local CLI installed.
What should I watch out for?Verification spends model calls on your accounts, and setup asks before installing any missing provider CLI. Each agent you add must be installed and signed in on its own. The roadmap still lists error recovery and reliability as work in progress.
Does it run on Windows, macOS and Linux?The README says it works across Windows, macOS and Linux. The AutoSwarm Desk desktop app is listed for Windows and macOS, while the CLI and IDE extensions are other ways to use it.
Is it free to use?The project is under the MIT license. The agents it connects, such as Claude Code or Codex, have their own accounts and model costs, which you pay for separately.
How does it decide which agent does what?It selects agents by their capabilities, connects complementary tools, runs parts of the task in parallel, and then verifies and merges the results into one coherent answer.
What does AutoSwarm Desk add over the CLI?AutoSwarm Desk is a desktop app for Windows and macOS. It gives you a private assistant and one window to start conversations with the supported agents, follow how they share the work and answer their approvals. It ships as one installer with Python included and updates automatically.
Do the agents share memory with each other?Yes. Each agent normally keeps its own context, and AutoSwarm is built to share context and memory across all of them, so you don't have to repeat information between tools.
Creator
HKUDS
Size
21.4 MB
Stars
2.6K
Age
4mo
Commits · 90 Days