Inter-office memo
- To
- Everyone running coding agents on Linux
- From
- Mo, Chief Executive
- Date
- November 20, 2026, 7:42 AM CST
- Re
- Parallel coding agents on Linux
The best app to run coding agents in parallel on Linux in 2026: 9 ways compared
On Linux, pick by your agents and your distribution. Enjoy runs every lab’s agent through usage limits, on x64; Orca and T3 Code are free and open source, on x64 and ARM64; Paseo suits a server you steer from your phone. For one lab, the ChatGPT app’s Linux preview runs Codex, and Claude’s beta runs Claude Code on Ubuntu or Debian. Conductor doesn’t run on Linux.
Disclosure: Enjoy is our product. We checked every app below against its maker’s own pages on November 20, 2026, and we say where another app is the better pick.
Ways to run coding agents in parallel on Linux, compared
Running agents in parallel means several sessions working at once, each on its own task, usually each in its own Git worktree so they don’t overwrite each other. The best-known app for it, Conductor, is for Macs: its installation page says “Conductor is not available for Windows or Linux yet.” It’s joining Meta, and its announcement now says “We’ll still be developing Conductor,” but nothing about Linux. Paradev, formerly Superset, offers “Linux as an experimental x64 AppImage,” and says “macOS is the primary target.” These are the ways that work on Linux today.
| Option | Best for | Price, checked November 20, 2026 | Linux packages | Processors | Agents |
|---|---|---|---|---|---|
| Enjoy | Every lab’s agent, across all your projects | A monthly subscription, free for 7 days, with your phone and one computer (pricing) | AppImage, .deb, .rpm | x64 | Claude Code, Codex, Grok Build, Antigravity, Muse Code, OpenCode |
| Orca | The most agents, free and open source | Free | AppImage, .deb, .rpm, AUR | x64, ARM64 | 40 set up, and any command-line agent |
| T3 Code | A free, open-source app with phone apps | Free | AppImage, .deb, AUR, Nix | x64, ARM64 | Claude Code, Codex, Cursor, OpenCode, Grok Build, Antigravity, Kiro, Pi, Muse Code, fx, and ACP agents |
| Paseo | A Linux server you steer from your phone | Free; hosted Hub from $0, Pro $15 a seat | AppImage, .deb, .rpm; npm for servers | x64 for the desktop app | Claude, Codex, OpenCode, Pi, OMP, Antigravity, Muse Code, and ACP agents |
| Emdash Classic | Every package format, on both processors | Free | AppImage, .deb, .rpm | x64, ARM64 | 35 command-line agents |
| Nimbalyst | A visual workspace around Claude Code and Codex | Free; Teams $20 a user, free in beta | .deb, AppImage | Not stated | Claude Code, Codex; OpenCode and Copilot in alpha |
| The ChatGPT app | Only Codex, in preview | Included in every ChatGPT plan, from Free; Plus is $20 | .deb, .rpm, an Arch script | x64, ARM64 | Codex |
| Claude’s desktop app | Only Claude, in beta | Included in Claude Pro ($20), Max ($100), Ultra ($200), Team and Enterprise | apt or .deb, Debian-based only | x86_64, arm64 | Claude Code |
Terminals and git worktree |
Free, if you live in a terminal | Free | Your distribution’s Git and tmux; Claude Squad and Agent Deck | Any | Any command-line agent |
Prices are per month as each maker lists them, before tax. The independent apps run your agents on your own Claude, ChatGPT or other plans and API keys, which those companies bill for separately.
Before you pick: your distribution and your processor
On Linux, the first question isn’t features. It’s whether an app ships a package for your distribution and your processor, and how it updates.
- Ubuntu and Debian get every app here, though the ChatGPT app lists only “Ubuntu 24.04 LTS and 26.04 LTS” and “Debian 13.” Claude’s desktop app is only for them: it needs “Ubuntu 22.04 or later, or Debian 12 or later,” and its Linux page says “only Debian-based distributions are supported today.”
- Fedora has .rpm packages from Enjoy, Orca, Paseo, Emdash Classic and the ChatGPT app, which lists “Fedora 43 and 44.”
- Arch has AUR packages for Orca (
stably-orca-bin) and T3 Code (t3code-bin), and an install script from OpenAI for the ChatGPT app. AppImages run there too. - ARM64 rules out Enjoy, which needs “a Linux computer with a 64-bit (x64) processor,” and Paseo’s desktop app, whose downloads for Linux are x86_64 only. Orca, T3 Code, Emdash Classic, the ChatGPT app and Claude’s desktop app ship ARM64 packages.
- Updates depend on the package. The AppImages from Enjoy and Orca update themselves. Enjoy’s .deb and .rpm don’t, and Orca’s “report available updates and give you the package-manager command to install them.” Claude’s and OpenAI’s apps update through their own package repositories.
Enjoy
Best for: running every lab’s agent side by side, across all your projects, without stopping when one plan runs out.
Enjoy runs Claude Code, Codex, Grok Build, Antigravity, Muse Code, and local models through OpenCode, on Linux, Mac and Windows, on the AI plans you already have. Each conversation runs its own agent, so several work at once, and an agent can work in a worktree when two tasks touch the same files. What it adds:
- Switch agents mid-conversation. When one agent is the wrong fit or out of usage, another picks up the same conversation.
- Keep going through limits. When an agent runs out of usage, Enjoy switches to another of your accounts that still has usage, or waits for the limit to reset, then continues. Saved accounts work for Claude Code, Codex and Grok Build.
- A superagent over everything. JOY keeps an eye on all your projects: ask it what needs you, catch up on what your agents did, or start work in any project.
On Linux, it installs as an AppImage, which updates itself, or as a .deb for Ubuntu and Debian or an .rpm for Fedora. While agents work, it asks your desktop’s screensaver, power manager and systemd to hold off sleep.
Price: one monthly subscription, and the first seven days are free, with a card to start. It includes your phone, any browser and one computer; each other computer is $9.99 a month, and each teammate you invite needs a subscription of their own. See pricing.
Watch for: Enjoy isn’t open source, and it’s a subscription on top of your AI plans once the free week ends. On Linux it’s for x64 only. The .deb and .rpm don’t update themselves, the AppImage “needs FUSE 2, xdg-utils, and desktop-file-utils,” and Enjoy “needs a keyring service such as GNOME Keyring installed and running.” Some graphics drivers, “most often NVIDIA’s,” leave its window blank until you start it with WEBKIT_DISABLE_DMABUF_RENDERER=1. Closing the Enjoy window quits it on Linux, so leave it open or minimized to stay reachable from your phone. Enjoy doesn’t stop two conversations from editing the same files, so overlapping work needs a worktree. Scheduled workflows run only while your computer is awake with Enjoy open. And it runs six agents, where Orca lists 40.
Orca
Best for: the widest choice of agents, free, with AppImage, .deb, .rpm and AUR packages.
Orca, from Stably AI, is “a desktop IDE for running multiple AI coding agents side by side,” and its docs say “Every task gets its own git worktree, its own agent terminal, and its own browser tab.” Its README says “Orca is free and open source under the MIT License.” For Linux, its install page offers an AppImage, a .deb and an .rpm, the README adds an AUR package, and its latest release has .rpm packages for x86_64 and aarch64. Its supported agents page lists 40 that “ship preconfigured in the built-in agent picker,” plus a Claude Agent Teams mode, and says “Orca works with any CLI agent.” A companion app for iOS and Android lets you “get pinged when an agent needs you, reply in chat.”
Watch for: “Choose the AppImage if you want Orca to update itself.” On Linux its command-line tool “is named orca-ide so it does not conflict with the GNOME Orca screen reader.” And it’s made for people who code: “If you’re looking for a no-code tool, Orca is not that.”
T3 Code
Best for: a free, open-source app that does one thing, running agents side by side, with phone apps.
T3 Code “is free, open source and fast,” under the MIT license. Its README says “It works with Claude Code, Codex, Cursor, OpenCode, Grok Build, Antigravity, Kiro, Pi, Muse Code, fx, and any ACP agent,” and “Each thread can work in its own worktree.” On Linux, its install guide says “AppImage and .deb downloads are available for x86_64 and ARM64,” with AUR packages for Arch and a Nix package. There are apps for iOS and Android and a web client, and its FAQ explains that “The phone is a remote. Your computer does the work and needs to stay on.”
Watch for: it sends “basic product usage events (provider, model, token totals, that kind of thing),” with “No prompts, no file contents,” until you set T3CODE_TELEMETRY_ENABLED=false. The .deb “asks for your password to install each update. If your desktop has no password prompt, the update fails.” And its FAQ warns that “Worktrees add up fast. One user cleared 500 GB.”
Paseo
Best for: agents on a Linux server or VM that you steer from your phone.
Paseo’s homepage says “Paseo is free and open source,” under Apache 2.0, with “custom implementations for Claude, Codex, OpenCode, Pi, OMP, Antigravity, and Muse Code,” and “many more via ACP.” Its README says “Run many agents at once, each in its own worktree,” on “iOS, Android, desktop, web, and CLI.” The desktop app comes as an AppImage, a .deb or an .rpm. For “a server, a VM, or any machine without the desktop app,” you install its command line with npm install -g @getpaseo/cli and start the daemon, and your phone, browser and desktop connect to it, through a relay that’s “end-to-end encrypted, Paseo can’t read your traffic,” or “directly over TCP, Tailscale, or another VPN.”
Watch for: its Linux desktop downloads are x86_64 only. Paseo Hub, for starting agents from GitHub, Slack and Discord, is free hosted for “50 agent runs a month” and one seat, or $15 a seat a month, and its FAQ says “access to the beta stays small and trusted.”
Emdash Classic
Best for: a free app with AppImage, .deb and .rpm builds for both x64 and ARM64.
Emdash is “a free, open-source agentic development environment,” under the Apache-2.0 license. Its README offers AppImage, DEB and RPM builds for “Linux x64” and “Linux ARM64,” and says “Keep every agent isolated in its own Git worktree and branch.” Its providers page says “Emdash supports 35 coding agent CLIs,” and the app detects Claude Code, Codex, Cursor, OpenCode, Amp, Devin and more once they’re installed.
Watch for: Classic gets fixes, not new features. The download page says “You’re downloading Emdash Classic. It stays available and keeps getting fixes, but new development has moved to Emdash 2.0: coding agents for your whole team, in the cloud.” Emdash 2.0 is in early access, and its pages give no launch date or price.
Nimbalyst
Best for: a visual workspace around Claude Code and Codex.
Nimbalyst is “The open-source visual workspace for building with Codex, Claude Code, and more,” under the MIT license, and it’s “Free for individuals.” Its README says “Run several coding agents at once, each isolated in its own git worktree,” and its features page says “Run Claude Code and Codex side by side, with OpenCode and GitHub Copilot available as alpha providers.” For Linux it offers a .deb for “Debian, Ubuntu, or a derivative” and an AppImage for other distributions. Its Teams plan is $20 a user a month, or $200 a user a year, “Free during Beta.”
Watch for: its release notes say “On Ubuntu 24.04 and later the AppImage needs an AppArmor profile before it will start,” so use the .deb there. There’s no .rpm, its pages don’t say which processors its Linux builds are for, and OpenCode and Copilot are still alpha providers.
The ChatGPT app, for Codex
Best for: people who only use OpenAI, on Ubuntu, Debian, Fedora or Arch.
“The ChatGPT desktop app for Linux is available in preview,” and OpenAI’s Linux page lists “Ubuntu 24.04 LTS and 26.04 LTS,” “Debian 13,” “Fedora 43 and 44” and “Arch Linux (current, fully updated rolling releases),” with “packages for x64 and ARM64 processors.” Installing it “configures a signed OpenAI package repository” for updates. The app lets you “Run projects in parallel,” and worktrees “let Codex run multiple independent chats in the same project without interfering with each other.” OpenAI’s pricing puts Codex on every ChatGPT plan from Free: Go is $8 a month, Plus $20 and Pro from $100.
Watch for: it runs only Codex. Computer Use “is available on macOS and Windows but not yet in the Linux preview,” “Other Linux distributions may work but aren’t formally supported,” and “Native Wayland support is experimental.” Free and Go get “GPT-6 Luna at Standard speed in the desktop app, subject to rollout.”
Claude’s desktop app, for Claude Code
Best for: people who only use Claude, on Ubuntu or Debian.
“Linux support for the Claude desktop app is in beta,” and Anthropic’s Linux page says it gives “the same Chat, Cowork, and Claude Code experience as on macOS and Windows: parallel sessions, visual diff review, an integrated terminal and editor, and live app preview.” It needs “Ubuntu 22.04 or later, or Debian 12 or later” on “x86_64 or arm64,” and installs from Anthropic’s apt repository. In a Git project, you can “turn on Worktree before you start a session to run it in its own Git worktree.” Claude Code needs a paid plan: Pro is $20 a month, or $17 billed yearly, Max is $100 and Ultra $200, and Team and Enterprise include it too.
Watch for: it runs only Claude, and “By default, Worktree is off and Claude works in the project folder itself.” On other distributions, “such as Fedora or Arch, run the CLI instead.” The Linux beta has no Computer Use and no dictation yet, and “The free claude.ai plan does not include Claude Code access.”
Terminals: git worktree, Claude Squad and Agent Deck
Best for: people at home in a terminal, on any distribution or processor, including over SSH.
You don’t need an app. Git lets one repository “support multiple working trees, allowing you to check out more than one branch at a time,” with git worktree add, and Claude Code makes one for you: run claude --worktree feature-auth, then, its docs say, “Run the same command with a different name in a second terminal to start an isolated parallel session.” Two free tools keep track of the sessions for you:
- Claude Squad, under AGPL-3.0, is “a terminal app that manages multiple Claude Code, Codex, Gemini (and other local agents including Aider) in separate workspaces.” It uses “tmux to create isolated terminal sessions for each agent” and “git worktrees to isolate codebases,” and its latest release has Linux builds for amd64 and arm64.
- Agent Deck, under the MIT license, “runs all your AI coding agents (Claude Code, Codex, Gemini CLI, OpenCode and more) in one terminal and shows which ones are working, waiting, or done.” It “Works on macOS, Linux and Windows (WSL).”
Watch for: you track every session, review every diff and clean up every worktree yourself. Claude Squad needs tmux and GitHub’s gh command installed first. And Git’s own manual warns that “Multiple checkout in general is still experimental, and the support for submodules is incomplete.”
When to stay with what you have
If you use one lab, its own app may be enough: on Ubuntu or Debian, Claude’s desktop app runs parallel sessions in worktrees, and the ChatGPT app runs parallel Codex chats on Ubuntu, Debian, Fedora and Arch, with no extra subscription. Both are still a beta and a preview on Linux, and if you’d rather not depend on either, claude --worktree and a terminal per task do the job for free.
Reach for another app when you use more than one lab’s agent, when you want every project’s agents in one window, when you want Claude Code in a window on Fedora or Arch, or when you want to keep working after one plan runs out.
How we compared
We read each maker’s homepage, downloads, install, pricing and docs pages, and their latest GitHub releases, on November 20, 2026, and every one is linked below. We quote their own words for the limits, and we don’t count features a maker hasn’t documented: where a maker’s pages don’t say which processors its Linux builds support, the table says so rather than guessing. This category changes weekly, so we recheck this guide against its sources at least once a month, and the date at the foot of the page says when we last did.
14.Questions.
What’s the best app to run coding agents in parallel on Linux?
- It depends on your agents and your distribution. Enjoy runs Claude Code, Codex and four other labs’ agents side by side on x64; Orca, T3 Code and Paseo are free and open source; and if you use one lab, the ChatGPT app’s Linux preview for Codex or Claude’s Linux beta for Claude Code is enough.
Is Conductor available for Linux?
- No. Conductor’s installation page says “Conductor is not available for Windows or Linux yet.” Paradev, formerly Superset, offers “Linux as an experimental x64 AppImage,” and says “macOS is the primary target.”
Is there a Claude Code or Codex desktop app for Linux?
- Yes, both, in early form. Claude’s desktop app “is in beta” on Linux, for “Ubuntu 22.04 or later, or Debian 12 or later”; the ChatGPT app, which runs Codex, “is available in preview” for Ubuntu, Debian, Fedora and Arch. Neither has Computer Use on Linux yet.
Which apps run on Fedora or Arch?
- On Fedora, Enjoy, Orca, Paseo, Emdash Classic and the ChatGPT app ship .rpm packages. On Arch, Orca and T3 Code have AUR packages, and OpenAI has an install script for the ChatGPT app. Claude’s desktop app supports “only Debian-based distributions” today, so Anthropic says to “run the CLI instead.”
Which apps run on ARM64 Linux?
- Orca, T3 Code, Emdash Classic, the ChatGPT app and Claude’s desktop app ship ARM64 builds, and Claude Squad has an arm64 release. Enjoy needs “a 64-bit (x64) processor,” and Paseo’s download page lists only x86_64 desktop builds for Linux.
Is there a free way to run coding agents in parallel on Linux?
- Yes. Orca, T3 Code, Paseo and Emdash Classic are free and open source, Nimbalyst is free for individuals, and Claude Squad or Agent Deck organize agents in a terminal for free. Your Claude, ChatGPT or other plans are billed separately in every case.