Optionalconfig: Config<EngineConfig>Optional Readonly Internal_Stores all configurations specific to editor instance.
ReadonlydocumentModel's document.
ReadonlymarkersModel's marker collection.
ReadonlyschemaModel's schema.
module:utils/observablemixin~Observable#decorate Decorated function for applying module:engine/model/operation/operation~Operation operations to the model.
This is a low-level way of changing the model. It is exposed for very specific use cases (like the undo feature).
Normally, to modify the model, you will want to use module:engine/model/writer~ModelWriter Writer.
See also {@glink framework/architecture/editing-engine#changing-the-model Changing the model} section
of the {@glink framework/architecture/editing-engine Editing architecture} guide.
The operation to apply.
Check whether given selectable is at a place in the model where it can be edited (returns true) or not (returns false).
Should be used instead of module:core/editor/editor~Editor#isReadOnly to check whether a user action can happen at given selectable. It may be decorated and used differently in different environment (e.g. multi-root editor can disable a particular root).
This method is decorated. Although this method accepts any parameter of Selectable type, the
~Model#event:canEditAt canEditAt event is fired with selectable normalized to an instance of
module:engine/model/selection~ModelSelection or module:engine/model/documentselection~ModelDocumentSelection
The change() method is the primary way of changing the model. You should use it to modify all document nodes
(including detached nodes – i.e. nodes not added to the module:engine/model/model~Model#document model document),
the module:engine/model/document~ModelDocument#selection document's selection, and
module:engine/model/model~Model#markers model markers.
model.change( writer => {
writer.insertText( 'foo', paragraph, 'end' );
} );
All changes inside the change block use the same module:engine/model/batch~Batch so they are combined into a single undo step.
model.change( writer => {
writer.insertText( 'foo', paragraph, 'end' ); // foo.
model.change( writer => {
writer.insertText( 'bar', paragraph, 'end' ); // foobar.
} );
writer.insertText( 'bom', paragraph, 'end' ); // foobarbom.
} );
The callback of the change() block is executed synchronously.
You can also return a value from the change block.
const img = model.change( writer => {
return writer.createElement( 'img' );
} );
The return type of the provided callback.
Callback function which may modify the model.
Creates a module:engine/model/batch~Batch instance.
Note: In most cases creating a batch instance is not necessary as they are created when using:
change(),enqueueChange().Optionaltype: BatchType
module:engine/model/batch~Batch#constructor The type of the batch.
Creates an operation instance from a JSON object (parsed JSON string).
This is an alias for module:engine/model/operation/operationfactory~OperationFactory.fromJSON OperationFactory.fromJSON().
Deserialized JSON object.
Creates a new position after the given module:engine/model/item~ModelItem model item.
Note: This method is also available as
module:engine/model/writer~ModelWriter#createPositionAfter Writer#createPositionAfter().
Item after which the position should be placed.
Creates position at the given location. The location can be specified as:
'end' (the position will be set at the end of that element),'before' or 'after'
(the position will be set before or after the given model item).This method is a shortcut to other factory methods such as:
createPositionBefore(),createPositionAfter().Note: This method is also available as
module:engine/model/writer~ModelWriter#createPositionAt Writer#createPositionAt(),
Optionaloffset: ModelPositionOffset
Offset or one of the flags. Used only when first parameter is a module:engine/model/item~ModelItem model item.
Creates a new position before the given module:engine/model/item~ModelItem model item.
Note: This method is also available as
module:engine/model/writer~ModelWriter#createPositionBefore Writer#createPositionBefore().
Item before which the position should be placed.
Creates a position from the given root and path in that root.
Note: This method is also available as
module:engine/model/writer~ModelWriter#createPositionFromPath Writer#createPositionFromPath().
Root of the position.
Position path. See module:engine/model/position~ModelPosition#path.
Optionalstickiness: ModelPositionStickiness
Position stickiness. See module:engine/model/position~ModelPositionStickiness.
Creates a range spanning from the start position to the end position.
Note: This method is also available as
module:engine/model/writer~ModelWriter#createRange Writer#createRange():
model.change( writer => {
const range = writer.createRange( start, end );
} );
Start position.
Optionalend: ModelPosition
End position. If not set, the range will be collapsed to the start position.
Creates a range inside the given element which starts before the first child of that element and ends after the last child of that element.
Note: This method is also available as
module:engine/model/writer~ModelWriter#createRangeIn Writer#createRangeIn():
model.change( writer => {
const range = writer.createRangeIn( paragraph );
} );
Element which is a parent for the range.
Creates a range that starts before the given module:engine/model/item~ModelItem model item and ends after it.
Note: This method is also available on writer instance as
module:engine/model/writer~ModelWriter#createRangeOn Writer.createRangeOn():
model.change( writer => {
const range = writer.createRangeOn( paragraph );
} );
Creates a new selection instance based on the given module:engine/model/selection~ModelSelectable selectable or creates an empty selection if no arguments were passed.
Note: This method is also available as
module:engine/model/writer~ModelWriter#createSelection Writer#createSelection().
// Creates selection at the given offset in the given element.
const paragraph = writer.createElement( 'paragraph' );
const selection = writer.createSelection( paragraph, offset );
// Creates a range inside an {@link module:engine/model/element~ModelElement element} which starts before the
// first child of that element and ends after the last child of that element.
const selection = writer.createSelection( paragraph, 'in' );
// Creates a range on an {@link module:engine/model/item~ModelItem item} which starts before the item and ends
// just after the item.
const selection = writer.createSelection( paragraph, 'on' );
// Additional options (`'backward'`) can be specified as the last argument.
// Creates backward selection.
const selection = writer.createSelection( element, 'in', { backward: true } );
See also: #createSelection:SELECTABLE createSelection( selectable, options ).
Optionaloptions: { backward?: boolean }Creates a new selection instance based on the given module:engine/model/selection~ModelSelectable selectable or creates an empty selection if no arguments were passed.
Note: This method is also available as
module:engine/model/writer~ModelWriter#createSelection Writer#createSelection().
// Creates empty selection without ranges.
const selection = writer.createSelection();
// Creates selection at the given range.
const range = writer.createRange( start, end );
const selection = writer.createSelection( range );
// Creates selection at the given ranges
const ranges = [ writer.createRange( start1, end2 ), writer.createRange( star2, end2 ) ];
const selection = writer.createSelection( ranges );
// Creates selection from the other selection.
// Note: It doesn't copies selection attributes.
const otherSelection = writer.createSelection();
const selection = writer.createSelection( otherSelection );
// Creates selection from the given document selection.
// Note: It doesn't copies selection attributes.
const documentSelection = model.document.selection;
const selection = writer.createSelection( documentSelection );
// Creates selection at the given position.
const position = writer.createPositionFromPath( root, path );
const selection = writer.createSelection( position );
// Additional options (`'backward'`) can be specified as the last argument.
// Creates backward selection.
const selection = writer.createSelection( range, { backward: true } );
See also: #createSelection:NODE_OFFSET createSelection( node, placeOrOffset, options ).
Optionalselectable: Optionaloptions: { backward?: boolean }Deletes content of the selection and merge siblings. The resulting selection is always collapsed.
Note: For the sake of predictability, the resulting selection should always be collapsed.
In cases where a feature wants to modify deleting behavior so selection isn't collapsed
(e.g. a table feature may want to keep row selection after pressing Backspace),
then that behavior should be implemented in the view's listener. At the same time, the table feature
will need to modify this method's behavior too, e.g. to "delete contents and then collapse
the selection inside the last selected cell" or "delete the row and collapse selection somewhere near".
That needs to be done in order to ensure that other features which use deleteContent() will work well with tables.
Selection of which the content should be deleted.
Optionaloptions: {Optionaldirection?: "forward" | "backward"The direction in which the content is being consumed. Deleting backward corresponds to using the Backspace key, while deleting content forward corresponds to the Shift+Backspace keystroke.
OptionaldoNotAutoparagraph?: booleanWhether to create a paragraph if after content deletion selection is moved to a place where text cannot be inserted.
For example <paragraph>x</paragraph>[<imageBlock src="foo.jpg"></imageBlock>] will become:
<paragraph>x</paragraph><paragraph>[]</paragraph> with the option disabled (doNotAutoparagraph == false)<paragraph>x[]</paragraph> with the option enabled (doNotAutoparagraph == true).Note: if there is no valid position for the selection, the paragraph will always be created:
[<imageBlock src="foo.jpg"></imageBlock>] -> <paragraph>[]</paragraph>.
OptionaldoNotResetEntireContent?: booleanWhether to skip replacing the entire content with a paragraph when the entire content was selected.
For example <heading1>[x</heading1><paragraph>y]</paragraph> will become:
<paragraph>^</paragraph> with the option disabled (doNotResetEntireContent == false)<heading1>^</heading1> with enabled (doNotResetEntireContent == true)OptionalleaveUnmerged?: booleanWhether to merge elements after removing the content of the selection.
For example <heading1>x[x</heading1><paragraph>y]y</paragraph> will become:
<heading1>x^y</heading1> with the option disabled (leaveUnmerged == false)<heading1>x^</heading1><paragraph>y</paragraph> with enabled (leaveUnmerged == true).Note: module:engine/model/schema~ModelSchema#isObject object and module:engine/model/schema~ModelSchema#isLimit limit elements will not be merged.
Removes all events listeners set by model instance and destroys module:engine/model/document~ModelDocument.
The enqueueChange() method performs similar task as the #change change() method, with two major differences.
First, the callback of enqueueChange() is executed when all other enqueued changes are done. It might be executed
immediately if it is not nested in any other change block, but if it is nested in another (enqueue)change block,
it will be delayed and executed after the outermost block.
model.change( writer => {
console.log( 1 );
model.enqueueChange( writer => {
console.log( 2 );
} );
console.log( 3 );
} ); // Will log: 1, 3, 2.
In addition to that, the changes enqueued with enqueueChange() will be converted separately from the changes
done in the outer change() block.
By default, a new batch with the default module:engine/model/batch~Batch#constructor batch type is created.
To define the module:engine/model/batch~Batch into which you want to add your changes,
use #enqueueChange:CUSTOM_BATCH enqueueChange( batchOrType, callback ).
Callback function which may modify the model.
The enqueueChange() method performs similar task as the #change change() method, with two major differences.
First, the callback of enqueueChange() is executed when all other enqueued changes are done. It might be executed
immediately if it is not nested in any other change block, but if it is nested in another (enqueue)change block,
it will be delayed and executed after the outermost block.
model.change( new Batch(), writer => {
console.log( 1 );
model.enqueueChange( new Batch(), writer => {
console.log( 2 );
} );
console.log( 3 );
} ); // Will log: 1, 3, 2.
In addition to that, the changes enqueued with enqueueChange() will be converted separately from the changes
done in the outer change() block.
Second, it lets you define the module:engine/model/batch~Batch into which you want to add your changes.
If you want to use default module:engine/model/batch~Batch#constructor batch type, use
#enqueueChange:DEFAULT_BATCH enqueueChange( callback ).
model.enqueueChange( { isUndoable: false }, writer => {
writer.insertText( 'foo', paragraph, 'end' );
} );
When using the enqueueChange() block you can also add some changes to the batch you used before.
model.enqueueChange( batch, writer => {
writer.insertText( 'foo', paragraph, 'end' );
} );
In order to make a nested enqueueChange() create a single undo step together with the changes done in the outer change()
block, you can obtain the batch instance from the module:engine/model/writer~ModelWriter#batch writer of the outer block.
A batch or a module:engine/model/batch~Batch#constructor batch type that should be used in the callback. If not defined, a new batch with the default type will be created.
Callback function which may modify the model.
Gets a clone of the selected content.
For example, for the following selection:
<paragraph>x</paragraph>
<blockQuote>
<paragraph>y</paragraph>
<heading1>fir[st</heading1>
</blockQuote>
<paragraph>se]cond</paragraph>
<paragraph>z</paragraph>
It will return a document fragment with such a content:
<blockQuote>
<heading1>st</heading1>
</blockQuote>
<paragraph>se</paragraph>
The selection of which content will be returned.
Checks whether given subject has any meaningful content.
Meaningful content is:
options.ignoreWhitespaces allows controlling whether this text node must also contain
any non-whitespace characters),This means that a range containing an empty <paragraph></paragraph> is not considered to have a meaningful content.
However, a range containing an <imageBlock></imageBlock> (which would normally be marked in the schema as an object element)
is considered non-empty.
Subject to check if includes meaningful content. Could be a model range, an element, a document fragment, or a selection.
Optionaloptions: { ignoreMarkers?: boolean; ignoreWhitespaces?: boolean }
OptionalignoreMarkers?: booleanWhether markers should be ignored.
OptionalignoreWhitespaces?: booleanWhether text node with whitespaces only should be considered empty.
Inserts content at the position in the editor specified by the selection, as one would expect the paste functionality to work.
Note: If you want to insert an {@glink framework/deep-dive/schema#object-elements object element} (e.g. a module:widget/utils~toWidget widget), see #insertObject instead.
This is a high-level method. It takes the #schema schema into consideration when inserting the content, clears the given selection's content before inserting nodes and moves the selection to its target position at the end of the process. It can split elements, merge them, wrap bare text nodes with paragraphs, etc. – just like the pasting feature should do.
For lower-level methods see module:engine/model/writer~ModelWriter Writer.
This method, unlike module:engine/model/writer~ModelWriter Writer's methods, does not have to be used
inside a #change change() block.
Inserting elements and text nodes into the model is not enough to make CKEditor 5 render that content
to the user. CKEditor 5 implements a model-view-controller architecture and what model.insertContent() does
is only adding nodes to the model. Additionally, you need to define
{@glink framework/architecture/editing-engine#conversion converters} between the model and view
and define those nodes in the {@glink framework/architecture/editing-engine#schema schema}.
So, while this method may seem similar to CKEditor 4 editor.insertHtml() (in fact, both methods
are used for paste-like content insertion), the CKEditor 5 method cannot be use to insert arbitrary HTML
unless converters are defined for all elements and attributes in that HTML.
Using insertContent() with a manually created model structure:
// Let's create a document fragment containing such content as:
//
// <paragraph>foo</paragraph>
// <blockQuote>
// <paragraph>bar</paragraph>
// </blockQuote>
const docFrag = editor.model.change( writer => {
const p1 = writer.createElement( 'paragraph' );
const p2 = writer.createElement( 'paragraph' );
const blockQuote = writer.createElement( 'blockQuote' );
const docFrag = writer.createDocumentFragment();
writer.append( p1, docFrag );
writer.append( blockQuote, docFrag );
writer.append( p2, blockQuote );
writer.insertText( 'foo', p1 );
writer.insertText( 'bar', p2 );
return docFrag;
} );
// insertContent() does not have to be used in a change() block. It can, though,
// so this code could be moved to the callback defined above.
editor.model.insertContent( docFrag );
Using insertContent() with an HTML string converted to a model document fragment (similar to the pasting mechanism):
// You can create your own HtmlDataProcessor instance or use editor.data.processor
// if you have not overridden the default one (which is the HtmlDataProcessor instance).
const htmlDP = new HtmlDataProcessor( viewDocument );
// Convert an HTML string to a view document fragment:
const viewFragment = htmlDP.toView( htmlString );
// Convert the view document fragment to a model document fragment
// in the context of $root. This conversion takes the schema into
// account so if, for example, the view document fragment contained a bare text node,
// this text node cannot be a child of $root, so it will be automatically
// wrapped with a <paragraph>. You can define the context yourself (in the second parameter),
// and e.g. convert the content like it would happen in a <paragraph>.
// Note: The clipboard feature uses a custom context called $clipboardHolder
// which has a loosened schema.
const modelFragment = editor.data.toModel( viewFragment );
editor.model.insertContent( modelFragment );
By default this method will use the document selection but it can also be used with a position, range or selection instance.
// Insert text at the current document selection position.
editor.model.change( writer => {
editor.model.insertContent( writer.createText( 'x' ) );
} );
// Insert text at a given position - the document selection will not be modified.
editor.model.change( writer => {
editor.model.insertContent( writer.createText( 'x' ), doc.getRoot(), 2 );
// Which is a shorthand for:
editor.model.insertContent( writer.createText( 'x' ), writer.createPositionAt( doc.getRoot(), 2 ) );
} );
If you want the document selection to be moved to the inserted content, use the
module:engine/model/writer~ModelWriter#setSelection setSelection() method of the writer after inserting
the content:
editor.model.change( writer => {
const paragraph = writer.createElement( 'paragraph' );
// Insert an empty paragraph at the beginning of the root.
editor.model.insertContent( paragraph, writer.createPositionAt( editor.model.document.getRoot(), 0 ) );
// Move the document selection to the inserted paragraph.
writer.setSelection( paragraph, 'in' );
} );
If an instance of the module:engine/model/selection~ModelSelection model selection is passed as selectable,
the new content will be inserted at the passed selection (instead of document selection):
editor.model.change( writer => {
// Create a selection in a paragraph that will be used as a place of insertion.
const selection = writer.createSelection( paragraph, 'in' );
// Insert the new text at the created selection.
editor.model.insertContent( writer.createText( 'x' ), selection );
// insertContent() modifies the passed selection instance so it can be used to set the document selection.
// Note: This is not necessary when you passed the document selection to insertContent().
writer.setSelection( selection );
} );
The content to insert.
Optionalselectable: ModelSelectable
The selection into which the content should be inserted. If not provided the current model document selection will be used.
OptionalplaceOrOffset: ModelPlaceOrOffset
To be used when a model item was passed as selectable.
This param defines a position in relation to that item.
at the insertion position.
Inserts an {@glink framework/deep-dive/schema#object-elements object element} at a specific position in the editor content.
This is a high-level API:
selectable before inserting,Writer.Writer, this method does not have to be used inside
a #change change() block.model.insertObject() does
is only adding nodes to the model. Additionally, you need to define
{@glink framework/architecture/editing-engine#conversion converters} between the model and view
and define those nodes in the {@glink framework/architecture/editing-engine#schema schema}.Use the following code to insert an object at the current selection and keep the selection on the inserted element:
const rawHtmlEmbedElement = writer.createElement( 'rawHtml' );
model.insertObject( rawHtmlEmbedElement, null, null, {
setSelection: 'on'
} );
Use the following code to insert an object at the current selection and nudge the selection after the inserted object:
const pageBreakElement = writer.createElement( 'pageBreak' );
model.insertObject( pageBreakElement, null, null, {
setSelection: 'after'
} );
Use the following code to insert an object at the current selection and avoid splitting the content (non-destructive insertion):
const tableElement = writer.createElement( 'table' );
model.insertObject( tableElement, null, null, {
findOptimalPosition: 'auto'
} );
Use the following code to insert an object at the specific range (also: replace the content of the range):
const tableElement = writer.createElement( 'table' );
const range = model.createRangeOn( model.document.getRoot().getChild( 1 ) );
model.insertObject( tableElement, range );
An object to be inserted into the model document.
Optionalselectable: ModelSelectable
A selectable where the content should be inserted. If not specified, the current module:engine/model/document~ModelDocument#selection document selection will be used instead.
OptionalplaceOrOffset: ModelPlaceOrOffset
Specifies the exact place or offset for the insertion to take place, relative to selectable.
Optionaloptions: {Additional options.
OptionalfindOptimalPosition?: "before" | "after" | "auto"An option that, when set, adjusts the insertion position (relative to
selectable and placeOrOffset) so that the content of selectable is not split upon insertion (a.k.a. non-destructive insertion).
'auto', the algorithm will decide whether to insert the object before or after selectable to avoid content splitting.'before', the closest position before selectable will be used that will not result in content splitting.'after', the closest position after selectable will be used that will not result in content splitting.Note that this option only works for block objects. Inline objects are inserted into text and do not split blocks.
OptionalsetSelection?: "after" | "on"An option that, when set, moves the module:engine/model/document~ModelDocument#selection document selection after inserting the object.
'on', the document selection will be set on the inserted object.'after', the document selection will move to the closest text node after the inserted object. If there is no
such text node, a paragraph will be created and the document selection will be moved inside it.
at the insertion position.Modifies the selection. Currently, the supported modifications are:
options.direction with a step specified in options.unit.
Possible values for unit are:'character' (default) - moves selection by one user-perceived character. In most cases this means moving by one
character in String sense. However, unicode also defines "combing marks". These are special symbols, that combines
with a symbol before it ("base character") to create one user-perceived character. For example, q̣̇ is a normal
letter q with two "combining marks": upper dot (Ux0307) and lower dot (Ux0323). For most actions, i.e. extending
selection by one position, it is correct to include both "base character" and all of it's "combining marks". That is
why 'character' value is most natural and common method of modifying selection.'codePoint' - moves selection by one unicode code point. In contrary to, 'character' unit, this will insert
selection between "base character" and "combining mark", because "combining marks" have their own unicode code points.
However, for technical reasons, unicode code points with values above UxFFFF are represented in native String by
two characters, called "surrogate pairs". Halves of "surrogate pairs" have a meaning only when placed next to each other.
For example 𨭎 is represented in String by \uD862\uDF4E. Both \uD862 and \uDF4E do not have any meaning
outside the pair (are rendered as ? when alone). Position between them would be incorrect. In this case, selection
extension will include whole "surrogate pair".'word' - moves selection by a whole word.Note: if you extend a forward selection in a backward direction you will in fact shrink it.
The selection to modify.
Optionaloptions: {Optionaldirection?: "forward" | "backward"The direction in which the selection should be modified.
OptionaltreatEmojiAsSingleUnit?: booleanWhether multi-characer emoji sequences should be handled as single unit.
Optionalunit?: "character" | "codePoint" | "word"The unit by which selection should be modified.
Editor's data model. Read about the model in the {@glink framework/architecture/editing-engine engine architecture} guide.