Skip to content

Aloud 0.2.0

Latest
Latest

Choose a tag to compare

View all tags
@github-actions github-actions released this 04 Oct 05:28
· 1 commit to main since this release
v0.2.0
a2f01cc

A big update: a simpler menu, a faster engine, and no download at all.

Changed

  • Native menu instead of the popup. Click the menu bar icon for play/pause, stop, replay, voice, speed, hotkey and launch at login. The caption and word-by-word highlighting are gone. If a hotkey press finds no text, Aloud beeps and shows the reason at the top of the menu.
  • Faster, lighter engine. Kokoro now runs on the Neural Engine through FluidAudio's CoreML port, instead of on the GPU through MLX. On an M2, for a 420-character paragraph, audio now starts after about 0.3 seconds instead of about 1.2 seconds, and peak memory went from about 8.6 GB to 0.7 GB.
  • The voice model is built into the app. Aloud makes no network connections. The model is checked against SHA-256 checksums at every launch, and a damaged copy repairs itself from the app. The old ~340 MB download in ~/Library/Application Support/Aloud/Models is deleted automatically.
  • 20 US voices. The UK voices are removed, because the new engine pronounces English with US rules only. If you had chosen a UK voice, Aloud switches to Heart.
  • Choosing a voice in the menu plays a short sample.

Faster

  • A long first sentence is split at its first comma, so audio starts sooner. Short sentences are merged, so there is less work per read.
  • The audio output starts while the first sentence synthesizes, not after.
  • The audio engine stops when a read ends, so the output device does not stay awake.
  • In apps where Aloud copies the selection (Slack, VS Code and other Electron apps), it reads the copied text as soon as it arrives instead of always waiting 300 ms. A hung app can no longer stall the hotkey for about 6 seconds.

Install

  1. Download the .dmg below.
  2. Open it and drag Aloud into Applications.
  3. Open Aloud once and dismiss the warning (it isn't notarized). Then open System Settings → Privacy & Security, click Open Anyway, and confirm.
  4. Grant Accessibility access when asked. The first launch takes about 10 seconds to prepare the voices.

If you are upgrading, macOS may ask for Accessibility access again.

Requires macOS 15+ on Apple Silicon. The app is about 120 MB, because it includes the voice model.

Assets 3
Loading