SnipsetDocs
FeaturesPricingBlogDownload
Documentation

AI Integration

Smart Shorthand

Expand shorthand selections with a local Qwen model — offline, on-device.

Smart Shorthand expands an abbreviated selection (e.g. Indonesian shorthand) into polished text using a dedicated local model. It works entirely offline with no cloud calls.

Model

  • Default model: hf.co/belajarcarabelajar/snipset-qwen-1.5b-gguf (~941 MB)
  • Settings: Settings → AI & Models → Smart Shorthand — Trigger key -- (also %%), Selection Shortcut Ctrl+Shift+S (DEFAULT_SMART_SHORTHAND_SHORTCUT), Unload when idle (5s / Immediately / 30s / 1m / 5m / 30m), Memory usage limit 80%/70%/90%/Disabled.
  • Download via the Model Catalog one-click hub; progress is shown in the Smart Shorthand section.

How to Use

  1. Select shorthand text in any app (e.g. sy mau ct email klien --).
  2. Press Ctrl+Shift+S (or type the trigger -- depending on mode).
  3. Snipset runs normalize_selected_shorthand_cmd locally and replaces the selection with the expanded form.

Commands

  • normalize_selected_shorthand_cmd — invoked via selection shortcut or trigger key. Respects smart_shorthand_model, smart_shorthand_idle_timeout, smart_shorthand_ram_threshold preferences.
Memory: The Qwen model is large. If you are low on RAM, set **Memory usage limit** to 70% or **Unload when idle: Immediately** to free memory quickly after use.
← PreviousAI CompanionNext →Template Variables Overview