Ross ROSS = Recommend OSS · open-source software intelligence for agents

coder/xum

A desktop app for isolated, parallel agentic development observed · 2026-08-28

github.com/coder/xum · homepage · TypeScript · AGPL-3.0 (copyleft) observed · 2026-08-28

Health v2 · maintenance only

79/100

  • Activity 99
  • Release rhythm 85
  • Longevity 25
How is this computed?

round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-03. Adoption (stars, forks) is never an input.

  • gap_med: 2
  • age_days: 350
  • days_rel: 21
  • days_push: 7
  • n_releases_24m: 64

Full methodology

Adoption not part of the score

1985 stars · 135 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Xum (formerly Mux) is an open-source desktop application for running multiple AI coding agents in parallel, each in its own isolated workspace (local, git worktree, or SSH remote). It is LLM provider-agnostic with multi-model support, integrates with VS Code, Cursor, Zed, Neovim, and JetBrains via ACP, and offers a central UI for managing agent suites with rich markdown output.

Use cases

  • run multiple coding agents in parallel on isolated git worktrees
  • manage a suite of AI coding agents from one desktop UI
  • run long-running agent refactors in the background and resume after restarts
  • compare several model answers to the same coding task with Best of N
  • connect coding agents to editors like VS Code, Zed, Neovim, or JetBrains via ACP
  • use local LLMs with Ollama for offline coding assistance
  • review agent-generated code changes with integrated code review
  • execute agents on a remote server over SSH

When to choose

  • you want to parallelize AI coding work across isolated workspaces with git divergence visibility
  • you need provider-agnostic multi-model support including local models via Ollama
  • you prefer a desktop GUI with keyboard-driven agent management over a single terminal agent
  • you want editor integration (VS Code, Zed, Neovim, JetBrains) for the last mile of agent tasks
  • you run agents on remote servers and need SSH-based workspaces

When to avoid

  • you only need a simple single-agent CLI like Claude Code without workspace management
  • you require full Windows support - it is still in alpha
  • you need a fully managed commercial product with SLA rather than an AGPL-3.0 open-source tool
  • your workflow does not involve git or coding tasks

Facets

application · maturity active

agent-framework llm-inference chat-interface developer-tools version-control code-review editor terminal-ui large-language-models developer-tools version-control windows cross-platform editor-plugin cli coding-agent-multiplexer git-worktrees parallel-agents acp vscode-extension ollama openrouter ssh-workspaces provider-agnostic agpl-3.0 ai-agents automation macos linux desktop

7 sources

Member repositories

RepositoryRoleHealth v2
coder/xummain79

For agents

markdown · JSON · MCP: product_card(name="coder/xum")

Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem