What is an HTML-to-PDF API?

An HTML-to-PDF API is a web service that converts HTML markup or a URL into a PDF file through a single API call, rendering the page in a real browser engine and printing the result to PDF.

The concept, expanded

Before HTML-to-PDF APIs existed, generating a PDF from an application usually meant building the document programmatically — placing text, lines, and boxes at exact coordinates using a PDF library. That works, but it means learning a separate layout model instead of reusing the HTML and CSS a team already knows.

An HTML-to-PDF API flips that around: you write (or already have) a normal web page — HTML for structure, CSS for layout and styling — and send it to the API. The API renders that page the same way a browser would, then converts the rendered result into a PDF. The output looks like the web page did, including fonts, tables, images, and print-specific CSS such as page breaks and margins.

How it works

Most HTML-to-PDF APIs follow the same three steps under the hood:

1. Load the source. The API receives raw HTML, a URL to fetch and render, or a stored template combined with JSON data.
2. Render in a browser engine. A headless browser (most commonly Chromium) parses the HTML and CSS, lays out the page, applies fonts, and executes any client-side JavaScript, just as it would for a person viewing the page.
3. Print to PDF. The browser engine's built-in print-to-PDF capability converts the rendered page into a PDF file, honoring page size, margins, and page-break CSS rules, and the API returns that file to the caller.

A minimal request looks like this:

POST /api/v1/pdf
Authorization: Bearer dw_live_your_key
Content-Type: application/json

{
  "source": { "type": "html", "html": "<h1>Invoice #1024</h1>" },
  "options": { "format": "A4", "margin": "20mm" }
}

Where Docweave fits

Docweave is an HTML-to-PDF API: POST /api/v1/pdf accepts HTML, a URL, or a template + JSON, renders it with Chromium, and returns the PDF. It's billed per document rather than per page, so a long report and a short receipt cost the same to generate, and every request can carry an idempotencyKey so a retry never produces a duplicate document.

Docweave is also MCP-native: the open-source generate_pdf tool (npx @docweave/mcp) lets an MCP client — Claude, Cursor, or another agent — call the same rendering pipeline directly as a tool call, without wiring up HTTP requests by hand.

FAQ