A visibility layer for Claude Code on Windows — see what your AI agent does in real-time

Built on Electron + xterm.js. Forked from cmux.

Passive Claude Code integration wmux observes Claude Code without changing how it works. A CDP proxy on localhost:9222 lets Claude Code's native chrome-devtools-mcp plugin control the wmux browser panel directly. Auto-configured hooks in settings.json report agent and tool activity to the sidebar. Zero setup — everything is auto-injected on startup.
Live browser visibility When Claude Code browses the web using its native browser tools (chrome-devtools-mcp), every action is visible in the wmux browser panel in real-time. Navigate, click, type, take screenshots — Claude Code uses its own tools, wmux just shows what's happening. CDP proxy on localhost:9222 bridges the connection transparently.
Live agent visibility When Claude Code uses its built-in Agent tool, wmux detects it via auto-configured hooks and displays agent activity in the sidebar — how many agents are working, what tools they're using. No changes to Claude Code's behavior. The sidebar also supports explicit wmux agent spawn for visible terminal tabs with round-robin pane distribution.
Activity indicators Sidebar dots show what Claude Code is doing at a glance. Orange pulsing = working. Green = done. Red = interrupted (Ctrl+C). Auto-notifications fire when a command finishes or is interrupted — in-app notification + Windows toast + taskbar flash.
Notification center Panes get a blue ring and tabs light up when agents need your attention. Supports OSC 9/99/777, wmux notify CLI, and idle detection. Click the bell icon to see all pending notifications — jump to any with one click or Ctrl+Alt+N.
Saved sessions Save your workspace layout (splits, CWDs, browser URL, shell) and restore it instantly. Click 💾 in the sidebar footer to save, 📂 to load. On startup, wmux auto-loads your last session — no more cd into project dirs and re-splitting panes every time.
Clipboard image paste Copy a screenshot (Win+Shift+S, Print Screen, etc.) and press Ctrl+V in a wmux terminal. The image is saved to a temp file and the path is injected into the terminal. Claude Code can then read the image — like pasting on claude.ai.
Vertical + horizontal splits See all your sessions at a glance in a sidebar. Git branch, PR status, working directory, listening ports, agent count, and notification text. Double-click to rename. Right-click for colors and workspace management. Split horizontally and vertically.
First-launch tutorial Interactive onboarding walks you through workspaces, splits, tabs, browser, and notifications. Reopen anytime from the ? button in the title bar.

• Scriptable -- Named pipe server (\\.\pipe\wmux) with a JSON-RPC API. Create workspaces, split panes, send keystrokes, read terminal content, control the browser via CDP, and spawn sub-agent terminals programmatically.
• Windows native -- ConPTY for proper terminal emulation, Windows toast notifications, taskbar flash on alerts, native title bar overlay.
• Windows Terminal + Ghostty compatible -- Import your themes, fonts, and colors from Windows Terminal settings.json or ~/.config/ghostty/config. Ships with 450+ bundled Ghostty themes.
• GPU-accelerated -- xterm.js with WebGL rendering for smooth terminal output at any speed.

Download wmux-0.4.0-win-x64.zip from GitHub Releases, extract anywhere, and run wmux.exe. No installer, no code signing, no admin required.

git clone https://github.com/amirlehmam/wmux.git
cd wmux
npm install
npm run build:main
npm run dev

I run a lot of Claude Code sessions in parallel. On macOS there is cmux, and it is exactly what I needed -- vertical tabs with live metadata, notification rings when agents need attention, a scriptable browser, and a socket API for automation. But I work on Windows, and nothing like it existed.

Windows Terminal has tabs but no notification system. You have to manually check each tab to see if an agent finished or is waiting for input. tmux works in WSL but loses all Windows integration. Electron terminals exist but none focus on the AI agent workflow.

So I built wmux — a visibility layer for AI coding agents. It doesn't replace Claude Code or change how it works. It passively observes and shows you what's happening. A CDP proxy on localhost:9222 lets Claude Code's native browser tools control the wmux browser panel — you watch every page load, click, and form fill in real-time. Auto-configured hooks in settings.json report tool usage and agent activity to the sidebar. When a command finishes or is interrupted, the sidebar dot changes color and you get a notification.

The sidebar shows exactly what each agent is doing -- the git branch it is on, the PR it opened, the ports it is listening on, how many sub-agents it spawned, and whether it needs your attention. When an agent finishes a task or hits a question, the pane gets a blue notification ring, the sidebar badge increments, and a Windows toast fires.

Shell integration scripts inject themselves into PowerShell, CMD, and WSL sessions. They report CWD changes, git branch switches, shell state, and PR status back to the sidebar via a named pipe in real time.

On first launch, wmux auto-configures itself as an observer: it injects a minimal informational block into ~/.claude/CLAUDE.md (so Claude Code knows the user can see its browser activity), adds a PostToolUse hook to ~/.claude/settings.json (for agent/tool visibility in the sidebar), and starts a CDP proxy on localhost:9222 (so chrome-devtools-mcp controls the wmux browser panel natively). No API keys needed — everything runs through the user's existing Claude Code session. Claude Code's behavior is unchanged.

Everything is automatable through the wmux CLI or the named pipe directly. The protocol matches cmux, so tools built for one work with the other.

wmux automatically injects integration scripts into your shells:

• PowerShell -- Overrides the prompt function. Reports CWD, git branch, dirty state, and shell state (working/done/interrupted) via NamedPipeClientStream. Preexec hook via PSReadLine detects when commands start. Background job polls gh pr view every 45 seconds.
• CMD -- Embeds OSC 9 escape sequences in the PROMPT variable for CWD reporting. Git branch detected via filesystem watcher on .git/HEAD.
• WSL (Bash/Zsh) -- PROMPT_COMMAND / precmd + preexec hooks. Detects interrupts via exit code 130. Communicates via temp file bridge.

Environment variables available in all shells:

All shortcuts are rebindable via Settings (Ctrl+,).

The wmux CLI communicates with the running app over the named pipe.

wmux ping                          # Check if wmux is running
wmux notify "Build complete"       # Send a notification
wmux new-workspace --title "API"   # Create a workspace
wmux list-workspaces               # List all workspaces
wmux split --right                 # Split focused pane
wmux send "npm test"               # Send text to terminal
wmux send-key Enter --ctrl         # Send keystroke
wmux read-screen --lines 50        # Read terminal content

# Browser (CDP-powered)
wmux browser open http://localhost:3000
wmux browser snapshot              # Accessibility tree with @eN refs
wmux browser click @e5             # Click element by ref
wmux browser type @e3 "hello"      # Type into input by ref
wmux browser fill @e3 "value"      # Set input value directly
wmux browser screenshot            # Base64 PNG screenshot
wmux browser get-text @e2          # Get element text content
wmux browser eval "document.title" # Run JavaScript

# Sub-agent spawning
wmux agent spawn --cmd "claude --resume abc" --label "Research"
wmux agent spawn-batch --json '[{"cmd":"claude","label":"Agent 1"},{"cmd":"claude","label":"Agent 2"}]'
wmux agent list                    # List all agents
wmux agent status <agent-id>       # Check agent status
wmux agent kill <agent-id>         # Kill an agent

wmux tree                          # Workspace/pane/surface hierarchy

Connect to \\.\pipe\wmux for programmatic control. Two protocols supported:

V1 (text, used by shell integration):

report_pwd <surface_id> <path>
report_git_branch <surface_id> <branch> [dirty]
report_shell_state <surface_id> idle|running|interrupted
notify <surface_id> <text>
ping

V2 (JSON-RPC, used by CLI and automation):

{"method": "workspace.create", "params": {"title": "Agent 1"}}
{"method": "workspace.list", "params": {}}
{"method": "surface.send_text", "params": {"id": "surf-...", "text": "npm test\n"}}
{"method": "surface.read_text", "params": {"id": "surf-...", "lines": 50}}

// Browser control (CDP-powered)
{"method": "browser.navigate", "params": {"url": "http://localhost:3000"}}
{"method": "browser.snapshot", "params": {}}
{"method": "browser.click", "params": {"ref": "@e5"}}
{"method": "browser.type", "params": {"ref": "@e3", "text": "hello"}}
{"method": "browser.fill", "params": {"ref": "@e3", "value": "hello"}}
{"method": "browser.screenshot", "params": {"fullPage": true}}
{"method": "browser.get_text", "params": {"ref": "@e2"}}
{"method": "browser.eval", "params": {"js": "document.title"}}
{"method": "browser.batch", "params": {"commands": [...]}}

// Sub-agent spawning
{"method": "agent.spawn", "params": {"cmd": "claude --resume abc", "label": "Research"}}
{"method": "agent.spawn_batch", "params": {"agents": [...], "strategy": "distribute"}}
{"method": "agent.list", "params": {}}
{"method": "agent.status", "params": {"agentId": "agent-..."}}
{"method": "agent.kill", "params": {"agentId": "agent-..."}}

{"method": "system.tree", "params": {}}

On relaunch, wmux restores:

• Window position and size
• Workspace layout (titles, colors, pin state)
• Split pane structure (directions and ratios)
• Working directory per terminal
• Browser panel URLs
• Active workspace and pane selection

wmux does not restore live process state. Active Claude Code, tmux, or vim sessions are not resumed after restart. Shells are respawned fresh in the saved working directories.

wmux reads configuration from two sources:

1. Windows Terminal -- %LOCALAPPDATA%\Packages\Microsoft.WindowsTerminal_...\LocalState\settings.json
2. Ghostty -- ~/.config/ghostty/config

Import either via Settings > Terminal > Import. Extracts font family, font size, color scheme, and palette. Default theme is Monokai. 450+ Ghostty themes bundled.

Two-process Electron model. Main process manages PTY spawning (node-pty/ConPTY), named pipe server, CDP browser bridge, port scanning, git/PR polling, notifications, Claude Code context injection, session persistence, and multi-window lifecycle. Renderer process runs React/Zustand with xterm.js (WebGL), recursive split pane layout, and the sidebar.

src/
  main/               # Electron main process
  renderer/            # React app (sidebar, splits, terminals, browser)
  preload/             # contextBridge API
  cli/                 # wmux CLI tool
  shared/              # Types shared between main and renderer
  shell-integration/   # PowerShell, CMD, WSL scripts

wmux is a Windows reimplementation of cmux, the macOS terminal for multitasking. Same design, same socket protocol, same philosophy. Tools built for cmux's API work with wmux.

• GitHub Issues -- bug reports and feature requests
• GitHub Discussions -- questions and ideas

wmux is open source under AGPL-3.0-or-later.