A workflow framework of specialized AI agents that takes a feature from idea to tested code inside Claude Code, pausing for human review between each wave.
EMI-Group/genesis
Open-source Elixir system where AI agents coordinate to build, validate, and evolve software from a single high-level objective.
- Stars
- 312
- Forks
- 45
- Open issues
- 0
- Last commit
- yesterday
What is it?
Genesis takes a high-level objective, such as a new project or a change to an existing repository, and coordinates AI agents to carry it out. Manager agents break the goal into tasks and review results, while executor agents implement and validate candidate changes in isolated Git worktrees. Accepted results are kept in Git history and SQLite runtime data, so later work starts from what was already built.
The project ships as a desktop app for macOS, Linux, and Windows built with Tauri, with a Phoenix LiveView dashboard for watching the agent tree, inspecting tasks and model usage, and reviewing diffs. A headless mode runs the dashboard and runtime on Linux or macOS servers, and remote development connects a local dashboard to agents on a server over SSH. Users configure their own model provider and API credentials.
The core is written in Elixir on the Erlang/OTP runtime. The repository is licensed under AGPL-3.0 and is developed by the EvoX Group.
Key features
- 01Decomposes objectives into tasks delegated between manager and executor agents.
- 02Runs candidate changes in isolated Git worktrees.
- 03Stores accepted results in Git history and task data in SQLite.
- 04Provides a desktop dashboard to view the agent tree, tasks, model usage, and diffs.
- 05Offers a headless server mode on Linux and macOS.
- 06Supports remote execution on servers over encrypted SSH tunnels.
Use cases
- Adding a feature to an existing repository while keeping current behavior intact.
- Starting a new software project from a written description of the desired outcome.
- Running long development tasks on a server while monitoring them from a laptop.
Quick start
- 1
Clone the repository
git clone https://github.com/EMI-Group/genesis.git - 2
Enter the project directory
cd genesis - 3
Fetch Elixir dependencies
mix deps.get - 4
Start the server, then open http://localhost:4100
mix phx.server
Things to know
- Licensed under AGPL-3.0, a copyleft license that requires sharing modified source when served over a network.
- Agents need a configured model provider and API credentials.
- The headless server has no built-in authentication and listens on localhost by default, so remote access needs an SSH tunnel.
- Running from source requires Elixir 1.20.4 and Erlang/OTP 29.
In collections
Topics
FAQ
- What is genesis?
- Open-source Elixir system where AI agents coordinate to build, validate, and evolve software from a single high-level objective. Genesis takes a high-level objective, such as a new project or a change to an existing repository, and coordinates AI agents to carry it out. Manager agents break the goal into tasks and review results, while executor agents implement and validate candidate changes in isolated Git worktrees. Accepted results are kept in Git history and SQLite runtime data, so later work starts from what was already built.
- Is genesis free to use? What license does it have?
- genesis is open source under the AGPL-3.0 license. Check the license terms before using it in commercial products.
- Is genesis still maintained?
- Actively maintained. The last commit to genesis was on Nov 6, 2026.
- What can I use genesis for?
- Adding a feature to an existing repository while keeping current behavior intact. Starting a new software project from a written description of the desired outcome. Running long development tasks on a server while monitoring them from a laptop.
- How do I install genesis?
- Clone the repository: git clone https://github.com/EMI-Group/genesis.git Enter the project directory: cd genesis Fetch Elixir dependencies: mix deps.get Start the server, then open http://localhost:4100: mix phx.server
- Where can I get genesis?
- The source code and installation instructions are on GitHub: https://github.com/EMI-Group/genesis
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open-source agent runtime and terminal TUI for long-horizon research and file-based work, with ReAct and parallel Agent Team modes.
A zero-code platform that generates production-style AI agents from natural-language descriptions, with built-in tools, memory and orchestration.
A Microsoft-led course of 18 lessons teaching beginners how to build AI agents, with Python code samples using Microsoft Agent Framework and Foundry.
An agent harness for Claude Code and Codex that adds multi-agent swarms, self-learning memory, vector retrieval and cross-machine agent federation.
Open-source local-first workspace for building AI agents visually, connecting models and MCP tools, and running agents in chat or automations.