Rendering to pixels

Render pages to pixels — for thumbnails, page rails, and headless output.

In short three paths produce rasters. Viewer.renderPageThumbnail(index, { width }) is the easy one for page rails. For full control, ViewerSession renders a page to tightly-packed RGBA8 at a target width or DPI, and the read-only Inspector.renderPage(index, dpi) returns PNG bytes without any canvas at all — ideal for tests and tooling.

Thumbnails (the easy path)

const raster = await viewer.renderPageThumbnail(0, { width: 240 });
// raster: { width, height, rgba: Uint8Array }  — draw into an ImageData / canvas

Full rasters, headless

// RGBA at a target width, from the session:
const png = await session.render_page_to_bytes(0, 1600);

// Or PNG bytes from the read-only inspector (no canvas, no camera):
const inspector = new Inspector(idmlBytes);
const pngBytes = inspector.renderPage(0, 144); // 144 dpi

SessionRaster/RenderedRaster are { width, height, rgba } — write the rgba into an ImageData and putImageData, or encode to PNG yourself. The Inspector path already returns encoded PNG.

Members

Frequently asked questions

Thumbnail vs. session render — which should I use? renderPageThumbnail for UI rails (it reuses the viewer's device). The session's render_page_to_bytes/render_to_bytes when you need a specific DPI or are rendering outside a viewer.

Can I render without a visible canvas? Yes — the Inspector.renderPage(index, dpi) path needs no canvas and returns PNG bytes, which is what the test corpus uses.

Is the output color-managed? The raster is the renderer's RGBA output. Color handling follows the engine; see the renderer for what the pipeline does with color.