fastpdf2png 1.3.0

Ultra-fast PDF to PNG converter

_{Built for Miruiq — AI-powered data extraction from PDFs and documents.}

fastpdf2png

Fast PDF to PNG converter. SIMD-optimized PNG encoding, automatic grayscale detection, multi-process scaling. MIT licensed.

Install

pip install fastpdf2png

Or build from source:

git clone https://github.com/nataell95/fastpdf2png.git && cd fastpdf2png
bash scripts/build.sh

Usage

CLI

./build/fastpdf2png input.pdf page_%03d.png 300 4 -c 2

Python

import fastpdf2png

images = fastpdf2png.to_images("doc.pdf")        # list of PIL images
fastpdf2png.to_files("doc.pdf", "output/")        # save PNGs to disk
data   = fastpdf2png.to_bytes("doc.pdf")          # raw PNG bytes
n      = fastpdf2png.page_count("doc.pdf")        # page count

# Batch processing — keep PDFium loaded between calls
with fastpdf2png.Engine() as pdf:
    for path in my_pdfs:
        images = pdf.to_images(path, dpi=150)

Node.js

const pdf = require("fastpdf2png");

pdf.toFiles("doc.pdf", "output/", { dpi: 150 });
const buffers = pdf.toBuffers("doc.pdf");
const count = pdf.pageCount("doc.pdf");

// Batch processing
const engine = new pdf.Engine();
await engine.toFiles("doc.pdf", "output/");
engine.close();

Performance

How it works

Rendering

Google's PDFium (the engine inside Chromium) renders each page into a raw BGRA bitmap in memory. This gives us pixel-perfect output identical to what Chrome displays.

Grayscale detection

Before encoding, a SIMD-accelerated pass scans every pixel to check if R == G == B. Most document pages (text, tables, charts) are grayscale — detecting this lets us encode them as 8-bit PNG instead of 24-bit RGB, cutting data size by 66% with zero quality loss. On ARM this uses NEON vld4/vceq intrinsics; on x86 it uses SSE/AVX2.

PNG encoding

Instead of the standard zlib/libpng pipeline, we use a patched libdeflate with a modified hash-skip optimization that skips redundant hash table insertions for long matches (+45% throughput). The compressed data goes directly into a pre-allocated output buffer — the PNG header, IDAT chunk, and IEND trailer are assembled around it with zero intermediate copies. CRC32 checksums are computed using hardware-accelerated instructions (CRC32 on ARM, PCLMUL on x86).

Parallelism

PDFium is not thread-safe, so we use fork() to create isolated worker processes. Each worker shares a single atomic page counter via mmap'd shared memory — workers grab the next unprocessed page with fetch_add, render it, and write the PNG to disk. Copy-on-write semantics mean the PDFium library and document data are shared across workers without duplicating memory.

Thread-local pools

Each worker maintains thread-local memory pools for pixel buffers and compression scratch space. After the first page warms up the pools, subsequent pages require zero malloc/free calls in the hot path.

CLI reference

fastpdf2png <input.pdf> <output_%03d.png> [dpi] [workers] [-c level]
fastpdf2png --info <input.pdf>
fastpdf2png --daemon

Flag	Default	Description
dpi	300	Output resolution
workers	1	Parallel processes
-c 0/1/2	2	Compression: fast / medium / best
--info		Print page count
--daemon		Persistent mode (stdin commands)

Platforms

OS	Arch	SIMD
macOS	arm64	NEON
macOS	x86_64	AVX2, SSE4.1
Linux	x86_64	AVX2, SSE4.1
Linux	arm64	NEON
Windows	x86_64	AVX2, SSE4.1

License

MIT. See LICENSE and THIRD_PARTY_LICENSES.md.