DocPuffin is a document-generation API: POST a template id + JSON, get a pixel-stable PDF back.

The demo on the site is the actual engine, not a mock. Pick a template (EU invoice or receipt), edit the JSON, hit render, and it sends the request to a function running in Frankfurt and returns the PDF in about 35 ms. Nothing is stored. Payloads are processed entirely in memory.

Why build another PDF API?

• First, EU-based rendering by default, so your data doesn’t need to cross the Atlantic.

• Second, invoice fields are designed around workflows common in tools like Fortnox and Visma, including organisation numbers, VAT rates and payment references. It’s mainly a formatting convenience, not compliance advice.

• Third, it uses credit packs instead of a subscription-based usage meter.

The stack: pdf-lib for PDF generation, it’s still pre-launch. The render engine is already working, and that’s what you’re using in the demo. API keys, documentation and the template editor will be coming over the next few weeks.

I’m preselling founding credit packs for $39, which includes 5,000 renders.

The feedback I’d value most is whether the template/JSON structure makes sense, and which document types you’d actually want to generate.