Skip to main content
SuperDoc has two navigation approaches depending on your use case:
ApproachUse caseStability
scrollToElement(id)Navigate to any element by its IDCross-session (for DOCX-imported content)
PositionTrackerTrack nodes that move during editsWithin a single session
scrollToElement takes any element ID — paragraph, comment, or tracked change — and scrolls to it. Use doc.extract() to get all IDs at once, or query.match for targeted lookups. 
// Extract all content with stable IDs
const { blocks, comments } = editor.doc.extract();

// Navigate to any block
await superdoc.scrollToElement(blocks[0].nodeId);

// Navigate to a comment
await superdoc.scrollToElement(comments[0].entityId);
This is the approach to use for:
  • RAG citations — store nodeId alongside embeddings, navigate on click
  • Cross-references — link to specific paragraphs from external UI
  • Search results — scroll to the matching paragraph
  • Cross-session addressing — IDs from DOCX-imported content survive reloads
For the full extraction pattern, see content extraction for RAG. For the cross-session pattern, see cross-session block addressing.

​
Track nodes during edits

When users edit a document, stored positions can drift. Use PositionTracker so navigation targets stay stable within the current session. 
// 1) Find hyperlink nodes
const found = editor.doc.find({
  select: { type: 'node', nodeType: 'hyperlink', kind: 'inline' },
});

// 2) Track each found node
const links = found.items.map((item) => ({
  item,
  trackerId: editor.positionTracker.trackNode(item),
}));

// 3) Navigate later (for example, from a sidebar click)
function goToLink(link) {
  if (!link?.trackerId) return;
  editor.positionTracker.goToTracked(link.trackerId);
}

​
Best practices

  • Use scrollToElement when you have an element ID from doc.extract() or the Document API.
  • Use PositionTracker when you need to follow nodes that move during edits.
  • For cross-session use, store nodeId values (not sdBlockId — those regenerate on each open).
  • Handle missing targets gracefully — both APIs return false if the element no longer exists.