Skip to content

WheelHouse 1.0.1

Choose a tag to compare

@ghbaud ghbaud released this 16 Jul 17:42

WheelHouse 1.0.1

First release with the graphical installer: download WheelHouse-Setup.exe below, double-click it, and follow the wizard. The PowerShell one-liner in INSTALL.md still works and installs the same thing.

The download is unsigned in this release, so Windows SmartScreen shows a warning ("Windows protected your PC"). Click More info, then Run anyway. INSTALL.md documents this.

Added

  • Graphical installer: WheelHouse-Setup.exe, a click-through setup wizard that runs the PowerShell installer for you. Built and attached to each release automatically.
  • Installer AI setup step: choose whether to enable the AI text-correction and help features and which server they use. The installer writes the server settings into config.toml and stores the API key in your user environment (never in a file); uninstalling removes the stored key, and re-running the installer preserves an existing AI setup by default.
  • Installer options for unattended runs (-SttProvider, -AutoStart, -StartNow, machine-readable progress output) -- these are what the graphical wizard uses to drive the install without console prompts.
  • "pattern manager" is now accepted as a spoken trigger for the Pattern Manager, alongside "x-ray patterns".

Fixed

  • Speech-engine fallback now resolves to the local Parakeet engine instead of cloud Google STT when the configured provider is unavailable -- audio no longer leaves the machine unless you explicitly chose a cloud engine.
  • The AI API key is read from the environment, not from config.toml, so a shared or backed-up config file cannot leak it.
  • Removed a misleading warning about cloud AI endpoints whose URL does not end in /v1.
  • Dictation now uses the fast caret-position read it was designed to use, instead of always taking the slow fallback path.
  • Quieter logs: per-keystroke and per-word diagnostic lines no longer repeat at INFO level during normal dictation.
  • Corrected stale voice-command names and examples in the in-app help.

Full changelog: CHANGELOG.md