How Type Detection Works
When you type /note followed by any text, AI analyzes your input using a confidence scoring system. Each word and pattern contributes points toward a specific note type:
| Type | Color | Trigger Keywords |
|---|---|---|
| Checklist | Peach | List items, action verbs, task-related words |
| Bug | Pink | ”crash”, “broken”, “error”, “fails”, “bug” |
| Idea | Blue | ”maybe”, “what if”, “explore”, “idea” |
| Snippet | Purple | Triple backticks, code keywords, operators |
| Link | Yellow | URLs, domain names |
| General | Yellow | Default fallback |
A minimum 2-point confidence threshold is required to trigger a specific type. If no type scores high enough, the note defaults to a general yellow note.
Tie-Breaking Priority
When multiple types score equally, the system uses this priority order:
- Checklist (most actionable)
- Bug (most urgent)
- Idea (exploratory)
- Snippet (reference)
- Link (reference)
Auto Title Generation
- Notes longer than 20 words get an AI-generated title as an H1 heading
- Short notes (under 20 words) skip the heading entirely — the content speaks for itself
When to Use /note vs Specific Commands
Use /note when you want zero friction — just dump your thought and let AI figure out the format. Use specific commands (/note-todo, /note-bug, /note-snippet) when you want guaranteed formatting and additional features like priority markers or git context.