A Conductor alternative for Windows

Conductor doesn’t run on Windows. It is a macOS app (installation), and since November 4, 2026, when it said it is joining Meta, it no longer takes new users (announcement). If you’re on Windows and want what Conductor does, Claude Code and Codex working side by side in separate worktrees, several apps do a similar job there today: Plyloft, Orca, T3 Code, Emdash, Paseo, Nimbalyst, Cline Kanban and Warp, each with a different idea of the job.

I build Plyloft on a Mac and test every release on Windows, because I won’t ship a feature to one and not the other. So this question comes up a lot for me.

Updated on November 9, 2026: Conductor said on November 4 that it is joining Meta, and Conductor Cloud shuts down on November 20. The section about Conductor is rewritten, with its dates and how to keep your cloud work, Orca and T3 Code are added, and the other apps were checked again.

What Conductor does, and why people look for an alternative

The news first, because it changes who is looking. On November 4, 2026, Conductor said it is joining Meta (announcement, changelog). From that day new users can no longer sign up or download the app. Conductor Cloud goes read-only on November 13 and shuts down on November 20, the day Conductor begins deleting cloud user data. The local Mac app keeps working and gets new models and critical bug fixes until early 2027; its team says it will have “something new” to use early next year, and that it can’t share much about it yet.

If you have cloud workspaces there, that part has a deadline, whatever you choose next. Conductor gives two ways to keep the work before November 20: right-click a cloud workspace and choose Transfer to local, which moves it and its chats to your Mac, or go to Settings → Account → Export data to download all your data.

Here is what the app does, for the people who have it.

Conductor is a Mac app where “each task gets its own workspace, branch, files, terminal, diff, and review path”, each workspace in its own git worktree, for Claude Code, Codex, Cursor and OpenCode (docs, worktrees). It runs Claude Code on your Pro or Max plan, on API billing or on another provider (Claude Code harness), and Codex with your ChatGPT sign-in, since September 29, 2026, or an API key (Codex harness).

It shows your CI and status checks (checks) and has forwarded failing checks to Claude since 0.12.0. It automerges pull requests once CI passes, supports Graphite stacks, and has a Review action where an agent inspects the current diff (review and merge, changelog). Conductor Cloud arrived on July 30, 2026, and the iOS app on October 2, 2026.

On price, as of October 22, 2026: Free ran local workspaces on your Mac; Pro was $50 a month per person (cloud workspaces, multiplayer, the API, the mobile app); Teams was $60 per user a month. That pricing page is gone. The announcement says existing free users can’t upgrade, Pro users keep the Pro features for free, and active subscribers get a refund.

The reason Windows users looked elsewhere was always simple: it doesn’t run there. At its launch on Hacker News in July 2025, the answer to that was “Mac only (for now)”, and it stayed that way (Show HN). Mac users now have the same question, with more time to answer it: the app they have keeps getting fixes until early 2027.

The options on Windows

	Windows	Agents	Review by another model	Board	Price
Conductor	No	Claude Code, Codex, Cursor, OpenCode	Review action, an agent on the diff	Sidebar of workspaces	No new sign-ups since November 4, 2026; winding down
Orca	Yes, an x64 installer for Windows 10 and 11	Any CLI agent; 34 named	No; your comments on diff lines go back to the agent	GitHub Projects boards; otherwise a sidebar of worktrees	Free, open source
T3 Code	Yes, x64 and ARM64 installers	Claude Code, Codex, Cursor, Grok Build, OpenCode, Antigravity, fx	No; a model can write a guide to a pull request (beta)	No; a sidebar of threads	Free, open source
Emdash	Yes	35 CLIs	No; you compare agents’ diffs side by side	Not verified	Free, open source
Paseo	Yes	Claude Code, Codex, Copilot, OpenCode and more	As a skill: two agents, different vendors when possible	No; a sidebar by state	Free, open source
Nimbalyst	Yes	Claude Code, Codex; others in alpha	An AI review in its pull request mode	Yes, sessions on a kanban	Free for individuals
Cline Kanban	Yes, via Node	Claude Code, Codex, Cline CLI, OpenCode	No	Yes	Free, open source
Warp	Yes	Claude Code, Codex, Gemini CLI, OpenCode	No	No	Free; Build $20; Business $50/user
Plyloft	Windows 11, 64-bit (x64), same features as Mac	Claude Code, Codex	Yes, in rounds, findings on the diff	Yes, by what needs you	Free (1 card building at a time); Pro €9; Team €9/seat

Not on Windows, for completeness: Paradev, which was called Superset until October 26, 2026 (“Windows is planned with no date”, FAQ), Sculptor (Apple Silicon Macs and Linux, repo) and Claude Squad: its Windows build launches, but creating a session fails, in an open issue reported on 1.0.17 (#275). Vibe Kanban runs on Windows through npx, but the company behind it shut down on April 10, 2026; the project continues as a community effort, and its README now says it is sunsetting (shutdown post, repo).

Orca

Stably AI’s MIT desktop app, at version 1.4.254 on November 8, 2026 (repo). On Windows it’s a signed x64 installer for Windows 10 and 11, with PowerShell or CMD as the shell (download, install). Every task gets its own git worktree, with its own branch, files and agent terminals, and it works with any CLI agent on the subscriptions you already have; its README names 34 (worktrees). Review is yours: you leave comments on lines of the diff and send them to the agent as one batch (Annotate AI Diff). Its README and docs describe no built-in review by a second model. It commits, pushes and opens the pull request, and sends failing checks or review comments back to an agent (commit and push). Agents can also run on an SSH host or a remote Orca server, and an iOS and Android companion app is in beta (mobile). Free and open source, on Mac, Windows and Linux. If you’re coming from Conductor on a Mac, it has a guide for migrating from Conductor, which opens Conductor’s worktrees where they are.

T3 Code

An MIT “agent harness control surface”: a server on your machine with an Electron desktop app, a web app and iOS and Android apps, at version 0.1.7 on November 4, 2026 (repo). On Windows there are x64 and ARM64 installers, winget install T3Tools.T3Code, and the option to run agents inside WSL (install). It drives Claude Code, Codex, Cursor, Grok Build, OpenCode, Antigravity and fx on your own subscriptions, or on an API key. Claude runs through Anthropic’s Agent SDK, and its FAQ says this drains Claude’s 5-hour limit faster than Claude Code itself, a bug it has reported to Anthropic (FAQ). A thread works in the current checkout or in a new worktree (threads). It commits, pushes and creates the pull request, and lets you review and merge pull requests; review guides, in beta, have Claude or Codex write a tour of a pull request for the person reviewing it (source control). Its docs describe no step where a second model reviews the work. Free and open source, on Mac, Windows and Linux.

Emdash

An open-source (Apache-2.0) “Agentic Development Environment” from Y Combinator’s Winter 2026 batch, at version 1.2.11 since November 6, 2026 (repo). It runs 35 agent CLIs, lets you review the changes of several agents side by side (you judge them), brings in issues from Linear, Jira, GitHub, Notion, Asana and others, and has SSH, GitHub Actions checks and scheduled automations (docs, providers). Open source and free, on Mac, Windows and Linux. Its site now calls this app Emdash Classic and says it “still gets fixes, but new development has moved to Emdash 2.0”, a cloud-first version for teams that has a waitlist and no prices yet (site). If you want the widest choice of agents for nothing, start here, knowing that.

Paseo

A local server, the “daemon”, with desktop, web, CLI, iOS and Android clients, about 23,600 stars on GitHub (repo). It runs the official agent CLIs on your own subscriptions: Claude Code, Codex, Copilot, OpenCode and others. Two of its skills come close to a review by another model: /paseo-committee has two agents analyze a hard problem independently, preferring “different provider families when possible”, and /paseo-advisor gives a second opinion on a design or a diff without editing files (skills). Free and open source, Apache-2.0, on Mac, Windows and Linux (download).

Nimbalyst

The successor of Crystal, which was deprecated in February 2026 (Crystal). An MIT-licensed visual workspace with a kanban of sessions, worktrees, diff review where you accept or reject each change, pull request reviews and visual editors, on Mac, Windows and Linux, with an iOS companion (repo, pull request reviews). It runs Codex and Claude Code side by side, with OpenCode, GitHub Copilot and Gemini in alpha. The app is free for individuals; the Wiki Teams plan is $20 per user a month, free during its beta (pricing).

Cline Kanban

A local web app you start with npx kanban, a research preview under Apache-2.0 (docs, repo). Each card gets its own worktree and terminal, with Cline CLI, Claude Code, Codex or OpenCode; you click on lines of the diff to leave comments that go back to the agent; Commit and Open PR finish the job; and linked cards start when the one before them is done. Free, with no account. It runs through Node, Windows included, with a few open Windows issues. An open issue reports that the Cline agent, using Claude Code as its provider, times out when run from Kanban (#9999).

Warp

A terminal that, since April 2026, works with Claude Code, Codex, Gemini CLI and OpenCode: vertical tabs, worktrees in one click, notifications from any agent, and inline comments from its code review sent to the running agent (blog). It doesn’t advertise a board or a review by another model. Free; Build $20 a month ($18 billed yearly) with $20 of agent usage; Business $50 per user ($45 yearly) (pricing). On Mac, Windows and Linux (download).

Plyloft

A desktop app for Mac and for Windows 11, 64-bit (x64), with the same features on both in every release (Plyloft on Windows). It runs the official claude and codex CLIs in real terminals, on your own Claude and ChatGPT plans, and never asks for an API key.

Each card gets its own copy of your project, a git worktree. Before any code, Plan with two AIs has one AI write the plan and the other critique it. Then one builds; when you press Finish and check, your project’s checks run and the other reviews read-only, in rounds, with findings such as F1 · Serious — … under the lines they name. Try it starts your app from the card’s copy, and Ship it opens a pull request and merges it when its checks pass, or merges on your machine. The Board sorts cards into Needs you, Not started, Working, On hold and Done, and Terminals holds your plain Claude, Codex and PowerShell sessions (Terminals).

Free builds one card at a time. Pro, €9 a month or €86.40 a year, adds cards in parallel, Build with both and Release several cards; Merge when checks pass is on Free too. Team is €9 per seat a month, with a shared board on GitHub Issues and shared skills (pricing).

What it doesn’t have, compared with Conductor: no cloud workspaces, so when your computer is off nothing is running; no mobile app; no Linux; and two agents rather than four. On Free, one card builds at a time, and cards in parallel need Pro. Nothing is imported from Conductor either: your code and its branches come over because they are in git, and the chats stay in Conductor’s export. If you’re moving from it, coming from Conductor goes through what’s the same, what’s different and the steps.

Next to the free tools above, the gaps are just as plain. Orca, T3 Code and Emdash run more agents and run on Linux. Orca, T3 Code and Paseo have phone apps, and Orca and T3 Code can run your agents on another machine. They cost nothing, where cards in parallel on Plyloft need Pro. What Plyloft adds is what happens after the agent says it’s done: your project’s checks run and a second AI reviews the change, in rounds, as a step of every card before the pull request. The pages of Orca, T3 Code and Emdash describe no such step; Paseo has a second opinion as a skill you call, and Nimbalyst an AI review once the pull request exists. Then the card merges, Release… makes a new version from what has reached your main branch, and a team works through a private GitHub repository, with a shared board on GitHub Issues.

Which one fits

This is my view, from what each project’s own pages say and from what I look for myself: the same features on Windows as on the Mac, a second AI reviewing the work, and the agents running on my own subscriptions. I haven’t run the alternatives on Windows, so try the one or two that fit before you settle.

• The most agents, free → Emdash or Orca. In both the review is yours; no second model reviews the diffs.
• A second opinion from another vendor, on your own subscriptions, with a phone client → Paseo. The review is a skill you call, not a step in the flow.
• A visual workspace with a kanban of sessions and an iOS companion → Nimbalyst.
• A simple local kanban, free and with no account → Cline Kanban.
• Your agents inside a terminal you already like → Warp.
• Claude Code and Codex working together from plan to ship, the other AI reviewing in rounds, on your own plans, with the same features on Windows and Mac → Plyloft. It’s the one I build, so weigh that.

For setting up the CLIs themselves on Windows, see Claude Code on Windows and installing Codex CLI. For the wider field, the Mac-only tools and the ones that shut down included, see the AI coding agent orchestrators compared. For what all of these are built on, see git worktrees for AI agents.

Plyloft is independent and not affiliated with Anthropic, OpenAI, Conductor or Meta.

If you’re on Windows and want Claude Code and Codex on one board, download Plyloft for Windows and try it free.