#218218th of 218 in Coding agents and AI editors

hephaestus

Multi-agent orchestrator where agents spawn tasks across phases on a Kanban board

Stars
1.2k
License
custom license
Last commit
Nov 2025
Language
Python
Stars, last 29 weeks+64
22 Apr 2026: 1,124 stars29 Apr 2026: 1,126 stars6 May 2026: 1,128 stars13 May 2026: 1,129 stars20 May 2026: 1,133 stars27 May 2026: 1,134 stars3 Jun 2026: 1,137 stars10 Jun 2026: 1,139 stars17 Jun 2026: 1,139 stars24 Jun 2026: 1,141 stars1 Jul 2026: 1,145 stars8 Jul 2026: 1,155 stars15 Jul 2026: 1,165 stars22 Jul 2026: 1,166 stars29 Jul 2026: 1,168 stars5 Aug 2026: 1,169 stars12 Aug 2026: 1,170 stars19 Aug 2026: 1,172 stars26 Aug 2026: 1,173 stars2 Sept 2026: 1,174 stars9 Sept 2026: 1,176 stars16 Sept 2026: 1,177 stars23 Sept 2026: 1,178 stars30 Sept 2026: 1,179 stars7 Oct 2026: 1,179 stars14 Oct 2026: 1,181 stars21 Oct 2026: 1,182 stars28 Oct 2026: 1,183 stars4 Nov 2026: 1,186 stars11 Nov 2026: 1,188 stars

1,124 on 22 Apr 20261,188 on 11 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Hephaestus runs CLI coding agents (Claude Code, OpenCode, Droid, Codex) in isolated tmux sessions. You define phase types such as analysis, implementation and validation, and agents create new tickets in any phase as they discover work. A Guardian process monitors agents against phase instructions, and a web UI shows the Kanban board and dependency graph. Hephaestus Dev adds five preset software-development workflows. Status is alpha.

Who it is for: Engineers experimenting with multi-agent coding workflows across CLI agents

Strengths

  • Agents create tasks in any phase, so workflows branch without predefined instructions
  • Isolated tmux sessions per agent, coordinated by Kanban tickets with blocking dependencies
  • Works with Claude Code, OpenCode, Droid or Codex as the agent CLI
  • Ships five preset dev workflows: PRD-to-software, bug fix, repo indexing, features, docs

Weaknesses

  • Alpha status and no tagged release
  • AGPL-3.0 per the README, which restricts closed-source network use
  • Heavy setup: tmux, Docker for Qdrant, Node.js, git repo and several API keys
  • Setup validation script is macOS only

What it needs

  • no GPU
  • Docker + Compose
  • Needs tmux, Git, Docker, Qdrant, Node.js, npm, Claude Code, OpenCode, Droid or Codex CLI
  • Models: OpenAI, Anthropic, OpenRouter, Azure OpenAI, Google AI Studio

Also in Coding agents and AI editors

See all 218
Also in Coding agents and AI editors
RankProjectScore
1claw-codeRust CLI coding agent harness using Anthropic and OpenAI-compatible APIs194.8k stars+724 (+0.4%) in 30 daysMIT96 out of 100
2t3codeDesktop, web and mobile front end for controlling local coding agent CLIs Live demo ↗ (opens in a new tab)57.9k stars+30,427 (+111%) in 30 daysMIT95 out of 100
3codewhaleTerminal coding agent in Rust that works with 40+ model providers41.2k stars+331 (+0.8%) in 30 daysMIT94 out of 100
#4ArchonYAML workflow engine that runs AI coding agents through fixed dev steps23.9k starsMIT93 out of 100
#5piMinimal terminal coding agent extensible with skills and packages121.6k stars+7,287 (+6.4%) in 30 daysMIT92 out of 100