Operazioni sulle pagine

6 operazioni di nome page.*. Ognuna si esegue in un batch con apply_ops, dry_run_ops, la CLI o l’admin.

Aggiornato il26/10/2026

Operation	What it does	Permission	Confirmation
page.read	Read a page row: slug, title, status, frame and SEO fields.	site.read	none
page.create	Create a page (slug, title per locale, and optionally frame, the theme layout frame it renders in, parentId, pageType, SEO and members-only fields) with an empty tree.	site.structure	none
page.update	Patch a page's fields as { pageId, patch }: slug, title, frame (the theme layout frame it renders in, such as default, which reflows at the site breakpoints, or a scaled canvas such as scale-canvas; layout.update edits a frame), status, SEO, members-only access or meta.	site.edit	none
page.delete	Delete a page and its tree.	site.structure	destructive
page.forkTemplate	Fork a theme template into an editable site page row.	site.structure	none
page.setPassword	Protect a page with a password, change it, or remove it with password: null.	site.edit	none

page.read

Read a page row: slug, title, status, frame and SEO fields. Page validation and ownership checks apply.

permission: site.readreads onlyclass: read

Esempio di input

{
  "pageId": "6f1d2c3b-4a59-4e8d-9c7b-1a2b3c4d5e6f"
}

Input

Field	Type	Required	Notes
pageId	string	yes	At least 1 character.

Output

Field	Type	Required	Notes
page	object	yes
page.slug	string	yes	At least 1 character.
page.title	record of string	yes
page.frame	string	yes
page.dynamic	boolean	yes
page.position	integer	yes
page.parentId	string \| null	yes
page.requireAuth	boolean	yes
page.requiredRoles	array of string	no	At most 50 items.
page.meta	record of any JSON	yes
page.pageType	string	no	One of "static", "blog-index", "article", "brand-canvas", "template", "region", "dynamic", "component".
page.templateId	string \| null	no
page.collectionKey	string \| null	no
page.status	string	no	One of "draft", "published".
page.authorUserId	string \| null	no
page.authorName	string \| null	no
page.featuredImageId	string \| null	no
page.excerpt	record of string \| null	no
page.metaTitle	record of string \| null	no
page.metaDescription	record of string \| null	no
page.ogImageId	string \| null	no
page.noIndex	boolean	no
page.canonicalUrl	string \| null	no
page.id	string	yes
page.passwordProtected	boolean	no
page.publishedAt	string \| null	no
page.createdAt	string	yes
page.updatedAt	string	yes
page.tree	object	no
warnings	array of string	no

I campi più annidati sono nello schema di output che risponde discover_operations con outputSchemas: true.

Provala

yarn splash page.read '{"pageId":"6f1d2c3b-4a59-4e8d-9c7b-1a2b3c4d5e6f"}' --dry-run

page.create

Create a page (slug, title per locale, and optionally frame, the theme layout frame it renders in, parentId, pageType, SEO and members-only fields) with an empty tree. status defaults to draft; a draft page stays out of the site publish, so pass status: "published" for a page the next publication.publish should serve. Fill the tree with tree.commit in the same batch: read the new root id with tree.read through a $ref to this operation's page.id. A section's anchor is the node field anchor in that tree, not a page field, meta.anchor or props.anchor. A pageType: "dynamic" page names a collection in collectionKey and is the detail page of its items: it answers /<slug>/<item slug> in each locale, and its tree binds the item. Page validation and ownership checks apply.

permission: site.structurewrites: pagesclass: transactionundo: yes

Esempio di input

{
  "slug": "studio",
  "title": {
    "en-US": "Studio"
  },
  "status": "published"
}

Input

Field	Type	Required	Notes
slug	string	yes	At least 1 character.
title	record of string	yes
frame	string	no
dynamic	boolean	no
position	integer	no
parentId	string \| null	no
requireAuth	boolean	no
requiredRoles	array of string	no	At most 50 items.
meta	record of any JSON	no
pageType	string	no	One of "static", "blog-index", "article", "dynamic".
templateId	string \| null	no
collectionKey	string \| null	no
status	string	no	One of "draft", "published".
authorUserId	string \| null	no
authorName	string \| null	no
featuredImageId	string \| null	no
excerpt	record of string \| null	no
metaTitle	record of string \| null	no
metaDescription	record of string \| null	no
ogImageId	string \| null	no
noIndex	boolean	no
canonicalUrl	string \| null	no
publishedAt	string (date-time) \| null	no

Output

Field	Type	Required	Notes
page	object	yes
page.slug	string	yes	At least 1 character.
page.title	record of string	yes
page.frame	string	yes
page.dynamic	boolean	yes
page.position	integer	yes
page.parentId	string \| null	yes
page.requireAuth	boolean	yes
page.requiredRoles	array of string	no	At most 50 items.
page.meta	record of any JSON	yes
page.pageType	string	no	One of "static", "blog-index", "article", "brand-canvas", "template", "region", "dynamic", "component".
page.templateId	string \| null	no
page.collectionKey	string \| null	no
page.status	string	no	One of "draft", "published".
page.authorUserId	string \| null	no
page.authorName	string \| null	no
page.featuredImageId	string \| null	no
page.excerpt	record of string \| null	no
page.metaTitle	record of string \| null	no
page.metaDescription	record of string \| null	no
page.ogImageId	string \| null	no
page.noIndex	boolean	no
page.canonicalUrl	string \| null	no
page.id	string	yes
page.passwordProtected	boolean	no
page.publishedAt	string \| null	no
page.createdAt	string	yes
page.updatedAt	string	yes
page.tree	object	no
warnings	array of string	no

I campi più annidati sono nello schema di output che risponde discover_operations con outputSchemas: true.

Provala

yarn splash page.create '{"slug":"studio","title":{"en-US":"Studio"},"status":"published"}' --dry-run

page.update

Patch a page's fields as { pageId, patch }: slug, title, frame (the theme layout frame it renders in, such as default, which reflows at the site breakpoints, or a scaled canvas such as scale-canvas; layout.update edits a frame), status, SEO, members-only access or meta. patch.status: "published" marks a draft page for the next site publish. patch.meta replaces the page's whole meta, so read the page first (page.read) and send its meta back with your change. meta.layoutOverrides overrides the frame for this page alone, layer by layer, where the frame's overridePolicy allows it: background ({ color, image, overlay, texture }: the page's own background colour, as raw CSS such as var(--ds-color-bg-inverse)), header, footer, scroll and chrome. Page validation and ownership checks apply.

permission: site.editwrites: pagesclass: transactionundo: yes

Esempio di input

{
  "pageId": "6f1d2c3b-4a59-4e8d-9c7b-1a2b3c4d5e6f",
  "patch": {
    "title": {
      "en-US": "Studio and team"
    },
    "status": "published"
  }
}

Input

Field	Type	Required	Notes
pageId	string	yes	At least 1 character.
patch	object	yes
patch.slug	string	no	At least 1 character.
patch.title	record of string	no
patch.frame	string	no
patch.dynamic	boolean	no
patch.position	integer	no
patch.parentId	string \| null	no
patch.requireAuth	boolean	no
patch.requiredRoles	array of string	no	At most 50 items.
patch.meta	record of any JSON	no
patch.pageType	string	no	One of "static", "blog-index", "article", "brand-canvas", "template", "region", "dynamic", "component".
patch.templateId	string \| null	no
patch.collectionKey	string \| null	no
patch.status	string	no	One of "draft", "published".
patch.authorUserId	string \| null	no
patch.authorName	string \| null	no
patch.featuredImageId	string \| null	no
patch.excerpt	record of string \| null	no
patch.metaTitle	record of string \| null	no
patch.metaDescription	record of string \| null	no
patch.ogImageId	string \| null	no
patch.noIndex	boolean	no
patch.canonicalUrl	string \| null	no

Output

Field	Type	Required	Notes
page	object	yes
page.slug	string	yes	At least 1 character.
page.title	record of string	yes
page.frame	string	yes
page.dynamic	boolean	yes
page.position	integer	yes
page.parentId	string \| null	yes
page.requireAuth	boolean	yes
page.requiredRoles	array of string	no	At most 50 items.
page.meta	record of any JSON	yes
page.pageType	string	no	One of "static", "blog-index", "article", "brand-canvas", "template", "region", "dynamic", "component".
page.templateId	string \| null	no
page.collectionKey	string \| null	no
page.status	string	no	One of "draft", "published".
page.authorUserId	string \| null	no
page.authorName	string \| null	no
page.featuredImageId	string \| null	no
page.excerpt	record of string \| null	no
page.metaTitle	record of string \| null	no
page.metaDescription	record of string \| null	no
page.ogImageId	string \| null	no
page.noIndex	boolean	no
page.canonicalUrl	string \| null	no
page.id	string	yes
page.passwordProtected	boolean	no
page.publishedAt	string \| null	no
page.createdAt	string	yes
page.updatedAt	string	yes
page.tree	object	no
warnings	array of string	no

I campi più annidati sono nello schema di output che risponde discover_operations con outputSchemas: true.

Provala

yarn splash page.update '{"pageId":"6f1d2c3b-4a59-4e8d-9c7b-1a2b3c4d5e6f","patch":{"title":{"en-US":"Studio and team"},"status":"published"}}' --dry-run

page.delete

Delete a page and its tree. Page validation and ownership checks apply.

permission: site.structurewrites: pagesclass: transactionconfirm: destructiveundo: yes

Esempio di input

{
  "pageId": "6f1d2c3b-4a59-4e8d-9c7b-1a2b3c4d5e6f"
}

Input

Field	Type	Required	Notes
pageId	string	yes	At least 1 character.

Output

Field	Type	Required	Notes
success	true	yes

Vuole confirmed: ["destructive"] sul batch, che la forma della CLI per una sola operazione non imposta: mettila in un file di batch ed esegui yarn splash apply <file> --dry-run.

page.forkTemplate

Fork a theme template into an editable site page row. Page validation and ownership checks apply.

permission: site.structurewrites: pagesclass: transactionundo: yes

Esempio di input

{
  "templateId": "page"
}

Input

Field	Type	Required	Notes
templateId	string	yes	At least 1 character.

Output

Field	Type	Required	Notes
page	object	yes
page.slug	string	yes	At least 1 character.
page.title	record of string	yes
page.frame	string	yes
page.dynamic	boolean	yes
page.position	integer	yes
page.parentId	string \| null	yes
page.requireAuth	boolean	yes
page.requiredRoles	array of string	no	At most 50 items.
page.meta	record of any JSON	yes
page.pageType	string	no	One of "static", "blog-index", "article", "brand-canvas", "template", "region", "dynamic", "component".
page.templateId	string \| null	no
page.collectionKey	string \| null	no
page.status	string	no	One of "draft", "published".
page.authorUserId	string \| null	no
page.authorName	string \| null	no
page.featuredImageId	string \| null	no
page.excerpt	record of string \| null	no
page.metaTitle	record of string \| null	no
page.metaDescription	record of string \| null	no
page.ogImageId	string \| null	no
page.noIndex	boolean	no
page.canonicalUrl	string \| null	no
page.id	string	yes
page.passwordProtected	boolean	no
page.publishedAt	string \| null	no
page.createdAt	string	yes
page.updatedAt	string	yes
page.tree	object	no
warnings	array of string	no

I campi più annidati sono nello schema di output che risponde discover_operations con outputSchemas: true.

Provala

yarn splash page.forkTemplate '{"templateId":"page"}' --dry-run

page.setPassword

Protect a page with a password, change it, or remove it with password: null. A visitor sees the password form instead of the page until they enter it; members are asked too. The password is stored hashed and never answered: page answers say only passwordProtected. Like the other page fields it reaches the public site with the next publish. Run it alone in a batch: the input is redacted from the history and the audit log, and the change is not undone.

permission: site.editwrites: pagesclass: transactionundo: explicit recovery

Esempio di input

{
  "pageId": "6f1d2c3b-4a59-4e8d-9c7b-1a2b3c4d5e6f",
  "password": "press-preview-2026"
}

Input

Field	Type	Required	Notes
pageId	string	yes	At least 1 character.
password	string \| null	yes

Output

Field	Type	Required	Notes
page	object	yes
page.slug	string	yes	At least 1 character.
page.title	record of string	yes
page.frame	string	yes
page.dynamic	boolean	yes
page.position	integer	yes
page.parentId	string \| null	yes
page.requireAuth	boolean	yes
page.requiredRoles	array of string	no	At most 50 items.
page.meta	record of any JSON	yes
page.pageType	string	no	One of "static", "blog-index", "article", "brand-canvas", "template", "region", "dynamic", "component".
page.templateId	string \| null	no
page.collectionKey	string \| null	no
page.status	string	no	One of "draft", "published".
page.authorUserId	string \| null	no
page.authorName	string \| null	no
page.featuredImageId	string \| null	no
page.excerpt	record of string \| null	no
page.metaTitle	record of string \| null	no
page.metaDescription	record of string \| null	no
page.ogImageId	string \| null	no
page.noIndex	boolean	no
page.canonicalUrl	string \| null	no
page.id	string	yes
page.passwordProtected	boolean	no
page.publishedAt	string \| null	no
page.createdAt	string	yes
page.updatedAt	string	yes
page.tree	object	no
warnings	array of string	no

I campi più annidati sono nello schema di output che risponde discover_operations con outputSchemas: true.

Provala

yarn splash page.setPassword '{"pageId":"6f1d2c3b-4a59-4e8d-9c7b-1a2b3c4d5e6f","password":"press-preview-2026"}' --dry-run

Generato dal registro delle operazioni e dal catalogo degli strumenti MCP (sorgente b47cd6cff6a50703).

Qualcosa non va in questa pagina? Scrivi a docs@librettocms.com.