Convert Images to PDF Online Convert one or more images into a multi-page PDF document with layout options.

Under the hood, it uses a combination of browser APIs and manual file format construction to achieve its functionality. When an image is added via the FileDropzone component, it creates a URL object for that image using `URL.createObjectURL`, allowing it to be loaded into an `Image` element. The `generate` function then scales and centers each image on a canvas element, which is referenced by the `canvasRef`. This scaled image is converted to a JPEG stream using `Canvas.toBlob` with a quality setting of 0.92. Each JPEG stream is embedded in the PDF document as an XObject resource with DCTDecode compression filter.

The tool manually constructs the PDF binary format according to the PDF 1.4 specification, including the catalog, page tree, page objects, content streams, and cross-reference table. The `TextEncoder` API is used to encode strings into Uint8Arrays for inclusion in the PDF document. Content streams use the cm operator to position and scale images on the page, with coordinates specified in points (1/72 inch). Each page object references its corresponding XObject resource, which contains the JPEG stream. After generating all page objects and content streams, it creates a cross-reference table and trailer to complete the PDF structure.

This manual construction approach allows for fine-grained control over the output PDF document, including the ability to embed images with specific compression filters and scaling factors. By using `Canvas.toBlob` and `TextEncoder`, it avoids relying on external libraries or server-side processing, making it a self-contained solution for converting images to PDF documents in the browser. When generating the final PDF, it creates a Blob object from the concatenated Uint8Arrays and triggers a download via an anchor element, resulting in a properly formatted PDF document that can be opened by any compatible viewer.