Loom v0.7.4 is out for macOS, Linux and Windows

Platforms

First-class on Linux.

AppImage, deb, and rpm builds of the same app, with secrets in your system keyring and the same six-session Claude Code fleet. No flagship platform, no afterthought port.

Packages

Pick your format.

AppImage

One portable file from the releases page. Download it, mark it executable, and run. A good fit when you want Loom without touching your package manager.

deb

A native package for apt-based distributions, installed and removed the way everything else on your system is.

rpm

The same build for rpm-based distributions. All three formats come from the GitHub releases page, and the installer is about 13 MB.

Setup

From download to fleet in minutes.

Loom drives six real Claude Code CLI sessions in parallel on your existing Claude login, so the fleet needs no API key. Install the CLI, sign in once, then launch whichever package you chose and brief a mission.

terminal
# install Claude Code and sign in
$ curl -fsSL https://claude.ai/install.sh | bash
$ claude login
# then launch Loom and hand it a goal

Questions

Linux specifics.

Where do my keys live?

Conductor API keys are stored in the operating system's keyring only, never in plaintext config files. Beyond that, a secret-path deny-list guards reads and writes, outbound requests go through an SSRF-guarded Rust HTTP proxy, every file, git, and shell operation passes a workspace authorization registry, and there is no telemetry.

Is the Linux build the same app as macOS and Windows?

Yes. It is one Tauri 2 + Rust + React 19 codebase: the same Conductor planning a goal into a mission DAG, the same fleet of six parallel sessions, and the same workspace with a WebGL terminal, CodeMirror 6 editor with inline AI autocomplete, full git graph, auto-detected web preview, and the Design tab.

Can I keep the reasoning model local?

Yes. The Conductor is bring-your-own-key and also speaks to local LM Studio and Ollama, or any OpenAI-compatible endpoint you run yourself. The app is free and Apache-2.0, with the full source on GitHub.

Hand it the work.
Walk away.

macOS, Linux, and Windows. Around 13 MB. Free and open source.