AICHE processes your speech through several AI layers. Each can be toggled independently in Settings > Enhance.

Enhancement Toggles

Message Ready

Converts casual speech into clean, professional text. Removes filler words (um, uh, like, you know), fixes grammar, and smooths out sentence structure.

Best for: emails, messages, documents.

Make English

Translates any language to English. Speak in your native language, get English output. Supports 50+ transcription languages.

Clean Language

Filters out inappropriate language. Useful for work recordings or shared notes.

Content Organize

Structures your rambling thoughts into organized text. Four levels:

Off - raw transcription, no restructuring
Light - basic cleanup, keeps your natural flow
Medium - groups related ideas, adds paragraph breaks
Strong - full reorganization into structured sections

Software Development

Vocabulary profile optimized for technical terms, programming languages, framework names, and developer jargon. Toggle this on when dictating code-related content.

Personal Dictionary

Teach the AI your specific words - names, product names, technical terms, abbreviations.

Where: Settings > Dictionary (desktop only currently)
Limit: 50 entries per account
Fields: Word (required) + "sounds like" (optional phonetic hint)
How it works: Applied server-side during transcription, so changes work on all your devices instantly

Example: Add "Kubernetes" with sounds-like "koo-ber-net-eez" so the AI never writes "cooper net ease."

Mobile UI for Personal Dictionary is not yet available, but entries added from desktop apply everywhere.

Auto-Categorization

Every transcription is automatically sorted into one of four categories:

Work - meetings, tasks, professional content
Life - personal notes, reminders, daily stuff
Ideas - creative thoughts, brainstorming, concepts
Misc - fragments, gibberish, or uncategorizable content

This is AI-powered and happens automatically after transcription. You can override by saying activation words like "for work" or "I have an idea" during your recording.

On mobile, you can filter your feed by category. On Obsidian, notes can auto-sort into category subfolders.

Voice Code

Hands-free dictation mode designed for coding. Available on desktop only (Mac, Windows, Linux).

Toggle: Control + Option + . (Mac) or Ctrl + Alt + . (Windows/Linux)
Real-time transcription via WebSocket streaming
Voice commands for numbered suggestions in Cursor and VS Code
Floating overlay shows current state
Mute/unmute without disabling: Control + Option + E (Mac) or Ctrl + Alt + E (Linux)

See the dedicated Voice Code guide for details.

Tips

Stack enhancements for best results. "Message Ready" + "Content Organize" works well for meeting notes.
Add your most commonly mispronounced words to Personal Dictionary first.
Use "Make English" if you naturally switch between languages mid-thought.
Software Development mode helps even if you're just talking about tech, not writing code.