Reference · §04

Custom types

Define your own fragment type by dropping a YAML schema into <vault>/schemas/. parc validates fragments against the schema on save and on reindex; the search DSL knows about every registered type automatically.

# Schema format

A schema describes a single type. The file name (without extension) is the type name.

# .parc/schemas/meeting.yml
name: meeting
title: Meeting
description: A scheduled or completed meeting

fields:
  attendees:
    type: list
    item_type: string
    required: false

  scheduled_at:
    type: datetime
    required: true

  duration_minutes:
    type: integer
    required: false
    min: 5
    max: 480

  status:
    type: enum
    values: [scheduled, completed, cancelled]
    default: scheduled

  outcome:
    type: text
    required: false

template: |
  ## Agenda

  - 

  ## Notes

  
  ## Action items

  - [ ]

# Field types

Type	YAML notes	DSL filter
`string`	Single-line text	`field:value`
`text`	Multi-line text	Free-text only
`integer`	Whole number, supports `min`/`max`	`field:>=N`
`float`	Decimal number, supports `min`/`max`	`field:>=N.M`
`boolean`	`true` / `false`	`field:true`
`date`	ISO date (no time)	`field:>=2026-03-01`
`datetime`	ISO 8601 with time	`field:>=2026-03-01T10:00:00Z`
`enum`	One of `values:`	`field:value`
`list`	Array of `item_type:`	`field:contains:value`

Every field accepts:

required: true|false (default false)
default: <value> — applied when the field is omitted on creation
description: <text> — shown in parc schema show

# Templates

The optional template: block is the body parc seeds the editor with for parc new <type> (without an inline title). Use it to encode your standard structure for that type — agenda, notes, action items.

If you prefer to keep templates in their own files, drop them in <vault>/templates/<type>.md instead. The schema-level template: and the file template are merged: the schema block applies first, then the file template overrides any common sections.

# Adding a schema

parc schema add ./meeting.yml

schema add validates the file against parc's schema-of-schemas, copies it to <vault>/schemas/meeting.yml, and refreshes the type registry. After this:

parc new meeting "Weekly sync"
parc list meeting
parc search 'type:meeting status:scheduled'

Add an alias to the vault's config.yml#aliases if you want a one-letter shortcut:

aliases:
  m: meeting

# Removing a schema

parc schema remove meeting

This deletes <vault>/schemas/meeting.yml and unregisters the type. Existing fragments of that type stay on disk but become invalid until you either re-add the schema or change their type: field by hand.

Built-in schemas (note, todo, decision, risk, idea) cannot be removed — they live inside the parc binary. You can override them by dropping a file with the same name into <vault>/schemas/, but this is rarely a good idea.

# A worked custom type

# .parc/schemas/bookmark.yml
name: bookmark
title: Bookmark
description: A URL worth keeping

fields:
  url:
    type: string
    required: true
    description: Canonical URL

  source:
    type: enum
    values: [hn, lobsters, reddit, twitter, mastodon, manual]
    default: manual

  read:
    type: boolean
    default: false

  rating:
    type: integer
    min: 1
    max: 5
    required: false

template: |
  > URL goes here

  ## Why I saved it

  

  ## Notes

After parc schema add bookmark.yml:

parc new bookmark "Local-first software" \
  --field url=https://www.inkandswitch.com/local-first/ \
  --field source=hn

parc search 'type:bookmark read:false rating:>=4'