Accessing the Engine

The PDFViewer component is powered by a high-performance PDFium engine compiled to WebAssembly. For advanced use cases, you can access the engine directly.

⚠️

The engine is stateless. Operations performed directly on the engine (like adding annotations) will not update the viewer UI. The viewer’s state is managed by plugins.

Use the engine only for:

Read-only operations (metadata, text extraction)
Export/rendering to external targets (save to file, render to canvas)
Advanced operations not supported by plugins

For operations that should reflect in the UI (adding annotations, navigating pages, zooming), always use the appropriate plugin instead.

Getting the Engine Reference

Use the onready callback to access the plugin registry, then get the engine:

App.svelte


<script lang="ts">
  import {
    PDFViewer,
    type EmbedPdfContainer,
    type PluginRegistry,
    type DocumentManagerPlugin,
  } from '@embedpdf/svelte-pdf-viewer';
 
  let container = $state<EmbedPdfContainer | null>(null);
  let registry = $state<PluginRegistry | null>(null);
 
  const handleInit = (c: EmbedPdfContainer) => {
    container = c;
  };
 
  const handleReady = (r: PluginRegistry) => {
    registry = r;
  };
 
  const getDocumentMetadata = async () => {
    if (!registry) return;
 
    // Access the engine
    const engine = registry.getEngine();
 
    // Get the active document
    const documentManager = registry
      .getPlugin<DocumentManagerPlugin>('document-manager')
      ?.provides();
    const document = documentManager?.getActiveDocument();
 
    if (engine && document) {
      // Use engine methods directly
      const metadata = await engine.getMetadata(document).toPromise();
      console.log('Document metadata:', metadata);
    }
  };
</script>
 
<button onclick={getDocumentMetadata}>Get Metadata</button>
<PDFViewer 
  oninit={handleInit}
  onready={handleReady}
  config={{ src: 'https://snippet.embedpdf.com/ebook.pdf' }} 
/>

Interactive Example

Click the button to fetch document metadata directly from the PDFium engine:

Common Engine Methods

The engine provides methods for low-level PDF operations. These are primarily read-only operations that are safe to use without affecting the viewer state:

Method	Description
`getMetadata(document)`	Get document metadata (title, author, dates, etc.)
`renderPage(document, page, options?)`	Render a page to an image (for export, not display)
`getPageText(document, pageIndex)`	Extract text content from a page
`getPageAnnotations(document, pageIndex)`	Read annotations on a page
`saveAsCopy(document)`	Save the document to a new buffer

All engine methods return an Observable that can be converted to a Promise using .toPromise().

ℹ️

Need to modify the document? Use plugins instead:

Add annotations → Use the Annotation Plugin
Change zoom → Use the Zoom Plugin
Navigate pages → Use the Scroll Plugin

Plugins manage state and ensure the UI stays in sync with your changes.

Full Engine Documentation

The engine provides many more capabilities including text extraction, page rendering, and annotation management.

For complete documentation of all available methods and advanced usage patterns, see the @embedpdf/engines documentation.

Accessing the Engine

Getting the Engine Reference

Interactive Example

Common Engine Methods

Full Engine Documentation

Need Help?