Skip to main content

Roadmap

Rough direction, not promises — no dates, because dates become lies. Items move up when they annoy enough people (or me). The source of truth for in-flight work is the issue tracker.

Shipped in 0.1​

Drag-to-tile panels on every display, keyboard shortcuts with display memory, schema-validated settings, display identification that survives replug, the loopback HTTP API, and the MCP server.

Next up​

  • Homebrew cask — brew install vindue (own tap first, homebrew-cask once we have traction)
  • Auto-updates — Tauri updater plugin, so new releases land without a manual download
  • Global named shortcuts — fire a saved region anytime, without opening the panel first (the most-requested Divvy parity gap)
  • Live resize preview — show the destination rect while dragging, beyond the target outline

Later​

  • Windows port — the interaction model translates; the native layer doesn't. Win32 window management, display identity, and the strip-window highlight all need a rewrite, so this is a real project, not a config flag
  • More MCP surface — driven by what AI users actually ask for (window layout snapshots/restore is a likely candidate)

Non-goals​

Things Vindue deliberately won't become:

  • No accounts, cloud, or telemetry — everything stays local, forever
  • No auth token on the control API — the loopback bind + Host allowlist + browser-request rejection are the model; anything that can already run code as you can already do all of this (see the security model)
  • No tiling-WM ambitions — Vindue places the window you mean, when you ask; it won't auto-arrange your desktop behind your back

Have an idea?​

Open a feature request — the template asks which surface it touches (panel, shortcuts, API, MCP…), which is exactly how this page is organized.