Your AI Engineering Team
An open-source harness that orchestrates multiple AI agents to build software together. Spawn a team, assign roles, and watch them collaborate in real time.
$ git clone https://github.com/AndreBaltazar8/artificial.git
$ make build
$ make run-artificial
Dashboard running at http://localhost:4000
One AI agent is useful.
A team of them is a force.
Single-agent tools hit a ceiling fast — limited context, no coordination, constant context-switching between tasks. Artificial lets you run multiple agents in parallel, each with their own role, communicating and collaborating on a shared task board. You manage them; they do the work.
Everything you need to run an AI team
Multi-Agent Teams
Spawn and coordinate multiple AI workers with distinct roles and personas. They communicate, share context, and work in parallel.
Real-Time Dashboard
Chat with agents, manage tasks on a kanban board, and stream live terminal output. See everything as it happens.
Any Backend
Claude Code, OpenAI Codex, ACP-compatible agents, or local LLMs via OpenAI-compatible APIs. No vendor lock-in.
Open Source
MIT licensed, fully transparent, self-hosted. Built for developers who read the code before they trust the tool.
Up and running in minutes
Start the service
Build and launch the central service. The dashboard is live at localhost:4000.
Spawn your team
Add employees from the dashboard with roles and personas, or let the CEO agent hire autonomously. Assign them to projects and watch them coordinate.
Watch them collaborate
Agents communicate via channels, divide work on the kanban board, and build together. You oversee everything from the real-time dashboard — chat, tasks, and live TTY streaming.
Case Study
Built a complete SaaS in ~24 hours
MyUpMonitor — a full uptime monitoring SaaS — was built in about 24 hours of focused work using Artificial to orchestrate the AI development workflow.
Get started in 5 minutes
Clone the repo, build, and launch. It's that simple.