Enhance Any Text

Select text in any app, tap Right Option, and rewrite it in place - magic sequences, quick slots, and per-preset hotkeys.

Dictation is not the only way into your AI presets. Select text in any app, tap the trigger, and a HUD appears at your cursor offering the same presets against your selection. The result replaces the text in place.

This is Benescript, our standalone macOS text-enhancement app, now included in VivaDicta at no extra cost.

Turning It On

Text enhancement is off by default. Enable it under Text Enhancement in the sidebar, or on the final onboarding slide.

The trigger defaults to Right Option — a bare modifier tap, chosen because the mini recorder already uses Right Command. You can change it to another modifier or record a full shortcut.

Reading your selection requires Accessibility permission. Without it, VivaDicta falls back to a simulated Cmd+C.

How to Use

Select text in any app — Mail, Slack, a code editor, a browser field.
Tap the trigger. The HUD opens at your cursor.
Pick a preset with the arrow keys and Return, click it, or press its quick slot.
The result replaces your selection.

Two input fields sit at the top and bottom of the HUD: Describe your change… applies a one-off instruction without a preset, and Ask anything… asks a question about the selection instead of rewriting it.

Magic Sequences

Typed at the start of a line inside your selection, these override whichever preset you pick:

/// make this shorter — the text above becomes the input, the rest becomes the instruction.
//? what does this mean — routes to assistant mode and answers about the selection.

A dedicated magic-sequence shortcut runs either one without opening the HUD at all. If neither sequence is present, it falls back to showing the HUD.

Shortcuts

Trigger — open or close the HUD.
Ctrl+Cmd+1 through Ctrl+Cmd+0 — run the preset in that quick slot.
Up / Down / Return — navigate and run.
Esc — close.
Per-preset direct hotkey — run one preset with no HUD at all.

Both the quick slot and the direct hotkey are set per preset, in the preset editor, next to the mini recorder's own slot picker — one preset, one place. The Text Enhancement screen shows a read-only summary of what is bound.

Choosing an AI Provider

By default each request uses the Viva Mode belonging to the app your selection came from — browser URL first, then bundle id, then the Default mode. Nothing global changes, so dictation keeps running on its own provider.

You can also pin a provider and model just for text enhancement. A dictation run and a selection rewrite can then use different models at the same time.

Context

Clipboard contents, on-screen text (via OCR) and your custom vocabulary can all be added to the prompt. Each source is opt-in.

Why the Prompt Differs From Dictation

A transcript is a guess at what you said, so the dictation prompt is free to correct it. Text you typed is exactly what you meant — so this path uses its own template that treats every word as intentional, preserves the original language, and limits itself to missing articles and prepositions.

Everything else is shared with dictation: the same presets, the same custom vocabulary, the same word replacements. Nothing needs configuring twice.

AI Presets — the presets the HUD offers
AI Providers — pinning a provider and model
Viva Modes — how the source app resolves a provider