Do I need to host Puppeteer or headless Chrome for n8n PDFs?

No. Dokifly renders HTML to PDF in the API, so the n8n workflow only needs an HTTP Request node and a destination for the file.

Where should the HTML template live?

In n8n itself (a Set node, a previous HTTP response, or a CMS). Send the finished HTML string to Dokifly; you do not need a custom backend to assemble the document.

HTML to PDF in n8n without a backend

n8n can generate a PDF from HTML without you operating a browser farm. Post the finished HTML to Dokifly, receive a PDF, and pass that file to email, Drive, or the next node.

Why generate PDFs from n8n instead of a server?

Most PDF stacks assume a long-running process: Puppeteer, Playwright, or a queue worker watching HTML jobs. That is a backend. n8n already has the data, the template, and the destination. The missing piece is a render step that does not require you to keep Chrome online.

Dokifly is an HTML-to-PDF API built for that gap — n8n, Make, Bubble, FlutterFlow, and plain REST. You send HTML. You get a PDF.

What does the n8n workflow look like?

A minimal path is four nodes:

Trigger — Webhook, schedule, or the app that already has the record (Stripe, Airtable, Typeform).
Build HTML — a Set node, a Code node, or an HTTP request that returns a complete HTML document (styles inlined or linked to a public stylesheet).
HTTP Request — POST the HTML to Dokifly. Use the endpoint, headers, and body shape from the Dokifly docs. Authenticate with the credentials you created in the product.
Output — Gmail, Slack, Google Drive, or another HTTP Request. n8n should treat the response as a binary file so the next node receives a .pdf.

Keep CSS simple and in the document. Remote fonts and images need publicly reachable URLs; the renderer cannot log into your private CDN.

How should I pass HTML in the HTTP Request node?

Use a complete document, not a fragment:

<!DOCTYPE html>
<html>
  <head>
    <meta charset="utf-8" />
    <style>
      body { font-family: sans-serif; color: #0a0a0a; }
    </style>
  </head>
  <body>
    <h1>Invoice {{ $('Trigger').item.json.number }}</h1>
    <p>Amount due: {{ $('Trigger').item.json.total }}</p>
  </body>
</html>

Resolve expressions in n8n so the string you POST is already filled in. Do not invent extra Dokifly fields — follow the current request format in the docs.

What should I do with the PDF after it returns?

Common next steps:

Attach it to an email (receipt, contract, packing slip).
Save it to Drive or S3 with a stable name like invoice-1042.pdf.
POST it to another system that expects a file, not HTML.

If the run fails, inspect the HTTP status and response body in n8n. HTML that is not a full document, missing auth, or an unreachable image URL are the usual causes — not a missing “PDF server” on your side.

When is this the wrong tool?

If you need a pixel-perfect design system with private assets behind a VPN, render inside that network. If you need a no-code automation that turns HTML into a file and moves on, an HTTP Request node plus Dokifly is the whole backend.