#140140th of 236 in Coding agents and AI editors

kandev

Kanban board that runs coding agent CLIs in parallel git worktrees

Stars
1.1k
License
AGPL-3.0
Last commit
Nov 2026
Last release
Nov 2026
Language
Go
Stars, last 29 weeks+1,070
25 Apr 2026: 73 stars2 May 2026: 149 stars9 May 2026: 170 stars16 May 2026: 224 stars23 May 2026: 274 stars30 May 2026: 305 stars6 Jun 2026: 321 stars13 Jun 2026: 332 stars20 Jun 2026: 360 stars27 Jun 2026: 390 stars4 Jul 2026: 414 stars11 Jul 2026: 436 stars18 Jul 2026: 458 stars25 Jul 2026: 494 stars1 Aug 2026: 533 stars8 Aug 2026: 565 stars15 Aug 2026: 645 stars22 Aug 2026: 685 stars29 Aug 2026: 713 stars5 Sept 2026: 756 stars12 Sept 2026: 788 stars19 Sept 2026: 815 stars26 Sept 2026: 852 stars3 Oct 2026: 894 stars10 Oct 2026: 922 stars17 Oct 2026: 973 stars24 Oct 2026: 1,012 stars31 Oct 2026: 1,055 stars7 Nov 2026: 1,096 stars14 Nov 2026: 1,143 stars

73 on 25 Apr 20261,143 on 14 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Kandev is a Go server with an embedded web UI that manages tasks on kanban or pipeline views and runs coding agent CLIs (Claude Code, Codex, Copilot, Gemini CLI and others) against them in parallel. Each task gets its own git worktree, and output is reviewed in a workspace with an editor, terminal, browser preview and git changes panel. Agents can run locally, in Docker, over SSH, in Kubernetes Pods or on sprites.dev.

Who it is for: Developers running several coding agents in parallel on local or remote machines

Strengths

  • Supports 20+ agent CLIs, mostly via ACP, and can mix agents per workflow step
  • Git worktrees isolate concurrent tasks; multi-repo tasks get one worktree per repo
  • Executors include local, Docker, SSH, Kubernetes and sprites.dev, with reusable profiles
  • Native Go binary with embedded frontend and SQLite; installs via Homebrew, Scoop or npx

Weaknesses

  • AGPL-3.0 license may restrict use in proprietary or hosted products
  • Office mode (persistent agent teams) is experimental and off by default
  • Agent CLIs must be installed separately; Cursor integration requires Cursor Pro
  • Orchestrates other agents rather than being one, so needs their accounts and keys

What it needs

  • no GPU
  • Docker
  • Needs Agent CLIs (Claude Code, Codex, etc.), Docker (optional), Node.js (only for npx launch)
  • Models: Claude Code, Codex, GitHub Copilot, Gemini CLI, OpenCode

What changed

From the release notes. No flag means the note does not say so, not that the upgrade is safe.

Also in Coding agents and AI editors

See all 236
Also in Coding agents and AI editors
RankProjectScore
1claw-codeRust CLI coding agent harness using Anthropic and OpenAI-compatible APIs194.8k starsMIT96 out of 100
2t3codeDesktop, web and mobile front end for controlling local coding agent CLIs Live demo ↗ (opens in a new tab)64.4k stars+36,149 (+128%) in 30 daysMIT96 out of 100
3codewhaleTerminal coding agent in Rust that works with 40+ model providers41.2k starsMIT94 out of 100
#4gooseRust AI agent with desktop app, CLI and API, extended through MCP56.7k stars+1,476 (+2.7%) in 30 daysApache-2.093 out of 100
#5ArchonYAML workflow engine that runs AI coding agents through fixed dev steps23.9k starsMIT93 out of 100