GoPdfSuit: Scaling PDF Generation to 600 Documents Per Second

When I started building GoPdfSuit, I thought: “PDF is just a document format. How hard can it be?”

Chinmay Sawant developed GoPdfSuit to handle 1.5 million financial PDFs in approximately 45 minutes. The system achieves sub-millisecond response times for two-page reports by bypassing traditional layout engines and implementing custom binary structures for font subsetting and coordinate tracking.

Why This Matters

Generating high-volume, archival-compliant documents like PDF/A-4 requires deep integration with binary structures and memory management that high-level libraries often lack. GoPdfSuit demonstrates that solving the ‘No Flow’ layout problem and implementing custom ICC profiles can yield a 92% reduction in operational costs compared to distributed clusters. Technical reality demands byte-level precision for digital signatures and cross-reference tables, which standard web-to-PDF tools often fail to optimize.

Key Insights

• Fixed-Coordinate Layout: GoPdfSuit tracks the Y-cursor manually, using a PageManager to trigger breaks for variable row heights instead of relying on CSS-style reflow.
• Binary Font Subsetting: The engine parses TTF binary structures including cmap and glyf tables to embed only used characters, ensuring consistent cross-device rendering.
• Memory Efficiency: Aggressive use of sync.Pool for bytes.Buffer and zlib writers allows the system to process 600 PDFs per second while minimizing GC pressure.
• Digital Signature Precision: Implementing PKCS#7 requires calculating exact ByteRange offsets and building CMS SignedData structures with ASN.1 encoding.
• Object ID Reservation: Pre-reserving IDs for Catalog and Metadata objects prevents cross-reference (xref) table corruption during sequential file serialization.

Working Examples

Layout engine tracking Y-coordinates and manual page breaks.

func (pm *PageManager) CheckPageBreak(requiredHeight float64) bool { return pm.CurrentYPos-requiredHeight < pm.Margins.Bottom } func (pm *PageManager) AddNewPage() { nextPageID := 3 + len(pm.Pages); pm.Pages = append(pm.Pages, nextPageID); pm.CurrentPageIndex = len(pm.Pages) - 1; pm.CurrentYPos = pm.PageDimensions.Height - pm.Margins.Top; pm.ContentStreams = append(pm.ContentStreams, bytes.Buffer{}); pm.PageAnnots = append(pm.PageAnnots, []int{}); }

Using sync.Pool to reduce GC spikes during high-throughput PDF generation.

var pdfBufferPool = sync.Pool{ New: func() any { buf := new(bytes.Buffer); buf.Grow(64 * 1024); return buf }, } var scratchBufPool = sync.Pool{ New: func() any { buf := make([]byte, 0, 128); return &buf }, }

Pre-reserving Object IDs to prevent xref table offset corruption.

metadataObjectID := pageManager.NextObjectID; pageManager.NextObjectID++; structTreeRootID := pageManager.NextObjectID; pageManager.NextObjectID++; if template.Config.PDFACompliant { iccProfileObjectID = pageManager.NextObjectID; pageManager.NextObjectID++; outputIntentObjectID = pageManager.NextObjectID; pageManager.NextObjectID++; grayICCProfileObjID = pageManager.NextObjectID; pageManager.NextObjectID++; }

Practical Applications

• High-Volume Financial Reporting: Generating 1.5 million documents in 45 minutes; Pitfall: Generating font subsets before signature rendering leads to missing glyphs in the final file.
• Archival Compliance (PDF/A-4): Direct embedding of ICC profiles for long-term storage; Pitfall: Using forward sRGB gamma instead of linearization curves causes washed-out colors in Adobe Acrobat.
• Mathematical Typesetting: Building a custom Typst-based engine for rendering formulas; Pitfall: Relying on headless browsers for simple vector math increases latency and infrastructure overhead.

References:

• https://dev.to/chinmay-sawant/when-i-started-building-gopdfsuit-i-thought-pdf-is-just-a-document-format-how-hard-can-it-be-41lb