PortBayvsKanbots
These two are the closest cousins in the agent-board space: both put your AI coding agents on a Kanban board and let a card dispatch the work. Kanbots is a cross-platform Electron app that runs 11 agent CLIs, isolates every run in a git worktree, syncs with GitHub Issues and can import Sentry errors straight onto the board. PortBay is a native macOS app where the card runs inside a full local dev environment — managed PHP and Node, a per-project database, trusted HTTPS on a real .test domain and a one-click public tunnel — so the agent builds against a running app.
Which one is right for you
You want the full stack, open source.
You want the agent working in a real environment — a database it can query, a site it can load over HTTPS, email capture, a shareable tunnel — provisioned by the same app that runs the board. You want a native macOS tool that is also your everyday local dev environment, not a board that assumes your stack is already running.
It already fits your workflow.
You run Windows, you want ACP-compatible agents specifically, or your workflow is GitHub-Issues-first — Kanbots mirrors card moves as issue labels and pulls Sentry errors onto the board. Its pre-push guard and explicit promote step are a stricter containment model than PortBay's, if agents landing changes on a branch worries you.
Where PortBay and Kanbots actually differ
The table below has the row-by-row detail. These are the three shapes it cannot draw: the job you already use one of them for, the job only one of them does, and what survives the move.
PortBay and Kanbots overlap on the everyday job: a local site on a real domain, over HTTPS, with the services it needs. If that is all you need, either will do it.
Claude Code
Codex
CursorMove a card to To Do and PortBay launches Claude Code, Codex or Cursor against a project that is already running. Kanbots has no board to dispatch from.
Point PortBay at the folders you already have. They stay where they are, .test domains are re-issued on the first run, and databases are imported rather than rebuilt.
Feature by feature
Kanbots rows come from Kanbots’s own site and docs, on the date below. PortBay rows come from the app’s source. We mark partial support honestly — including where the other side wins.
Kanbots facts on this page were last checked against github.com/leodavinci1/kanbots on . Pricing and limits change — check their site before you decide.
Already trying Kanbots?
There's no board import between the two — but there doesn't need to be. Point PortBay at the same repository Kanbots has been branching worktrees from, press play, and the runtime, database and HTTPS come up around it. Dispatch the same Claude Code or Codex from a PortBay card and the agent works against the running app instead of a bare worktree. If you rely on Kanbots' GitHub Issues sync or Sentry triage, you can keep it for that flow and use PortBay where the agent needs a live environment.
Install PortBay and add the project folder you've been dispatching Kanbots agents at.
Press play — PortBay provisions the runtime, database and HTTPS for that project.
Create a card, assign Claude Code or Codex, move it to Todo, and the agent works against the running app.
PortBay vs Kanbots, in plain terms
Don’t see your question answered here? The docs go deeper — every feature, flag and failure mode.
Read the docsPortBay runs the environment the card executes in. The same app that dispatches Claude Code or Codex also provisions managed PHP/Node runtimes, a per-project MySQL or PostgreSQL database, trusted HTTPS on a real .test domain, email capture and a one-click Cloudflare tunnel. Kanbots' branch preview can start your existing dev script inside a worktree, but it provisions nothing — no runtimes, no database, no HTTPS, no tunnels.
The board concept is similar — both put agent tasks on Kanban cards with live activity. The difference is what sits underneath. Kanbots assumes your stack already runs and isolates agents in git worktrees; PortBay is also the local dev environment, so a card has a running application — database, HTTPS URL, email capture — behind it from the moment you press play.
The rosters are about the same size. Kanbots dispatches 11 agent CLIs — Claude Code, Codex, Gemini, Cursor CLI, GitHub Copilot CLI, Amp, OpenCode, Droid, CCR and Qwen — plus any ACP-compatible CLI. PortBay dispatches 11 out of the box — Claude Code, OpenAI Codex, Cursor, Gemini, Aider, Copilot CLI, OpenCode, Amp, Qwen Code, Antigravity and Ollama — plus any other CLI via a Custom command template. Kanbots covers Droid and CCR; PortBay covers Antigravity and local Ollama models. Both detect whichever you already have installed.
Kanbots is new: the repository shows a first release around May 22, 2026, MIT-licensed, with roughly 450 GitHub stars when we checked in June 2026. Its macOS and Windows builds were unsigned at that point, so installation needed a Gatekeeper or SmartScreen workaround, which their README documents. Check the repository for its current release cadence before you commit to it.
Choose PortBay when the agent needs a real running environment and you want one app that's also your daily local dev tool. Choose Kanbots when you're on Windows, you need ACP-compatible agents, or your triage flow lives in GitHub Issues and Sentry and you want the board to mirror it.
Give your projects and your agents a real local home.
Download for macOSFree & open source · macOS 11+ on Apple Silicon · Pro from $10/mo