I Built a PDF API That Doesn't Use Chrome – Here's Why

#webdev #api #saas #pdf

I Built a PDF API That Doesn't Use Chrome

Most PDF generation APIs spin up a headless browser (Puppeteer, Playwright) to render HTML into a PDF. It works, but it's slow, expensive, and unpredictable.

So I built DocForge – a PDF generation API that uses PDFKit directly for deterministic, fast output.

How It Works

POST /api/generate
{
  "template": "invoice",
  "data": {
    "invoice_number": "INV-001",
    "from": { "name": "Acme Corp" },
    "to": { "name": "Client Corp" },
    "items": [...],
    "total": 49.99
  }
}
→ Returns a PDF in ~3 seconds

Why Not Chrome?

	Headless Chrome	PDFKit (DocForge)
Speed	5-15 seconds	~3 seconds
Memory	500MB+ per instance	~50MB
Deterministic	❌ Different renders possible	✅ Same input = same output
Cost	Expensive at scale	Cheap at scale