@catlabtech/webcvt-image-pdf

Wrap an image into a one-page PDF — a clean-room PDF writer with zero runtime dependencies (no third-party PDF library).

• JPEG → PDF: embedded byte-for-byte via DCTDecode (lossless, no canvas).
• PNG / WebP / BMP / GIF → PDF: decoded to pixels and embedded as a Flate-compressed DeviceRGB image with a DeviceGray soft mask for transparency.

Installation

npm i @catlabtech/webcvt-image-pdf

Supported conversions

PDF is always the output:

Input	Output
JPEG	PDF (lossless DCTDecode)
PNG, WebP, BMP, GIF	PDF (Flate + alpha soft mask)

Non-JPEG sources use a canvas pixel-bridge (OffscreenCanvas). JPEG works anywhere.

Usage

import { registerPdfBackend } from '@catlabtech/webcvt-image-pdf';
registerPdfBackend();

import { convert } from '@catlabtech/webcvt-core';
const pdf = await convert(jpegBlob, { format: 'pdf' });

Free functions jpegToPdf(bytes) and imageDataToPdf(imageData) are also exported.

Notes

• Each image becomes one page; the page size equals the image dimensions in points.
• Security caps: 256 MiB max input, 25 MP max image. CMYK JPEGs are rejected in v1.
• No wasm and no wasm-unsafe-eval needed — pure TypeScript plus the platform CompressionStream for the Flate path.

Source

packages/image-pdf/src