← All @molecule/* packages · App templates

@molecule/api-pdf-puppeteer

Provider bond · pdf · API (Node) · v1.0.1 · Apache-2.0

Puppeteer HTML-to-PDF provider for molecule.dev

npm install @molecule/api-pdf-puppeteer

npm · Source on GitHub · Implements @molecule/api-pdf

How it works

@molecule/api-pdf-puppeteer is a provider bond on the API (Node) side: it implements the pdf core interface (@molecule/api-pdf) with a concrete vendor or library behind it.

Your code calls the core; you wire this provider once at startup. Swapping vendors later is one line in that wiring, not a rewrite.

import { setProvider } from '@molecule/api-pdf'
import { provider } from '@molecule/api-pdf-puppeteer'

setProvider(provider)

Works with: @molecule/api-pdf

Reference

Auto-generated, AI-first package reference for the molecule.dev ecosystem. It is written to be read by coding agents as much as by people, and is generated from this package's source — edit src/index.ts JSDoc, not this file.

Puppeteer PDF provider for molecule.dev.

High-fidelity HTML-to-PDF rendering powered by headless Chrome via Puppeteer, with PDF manipulation (merge, watermark, metadata) via pdf-lib.

Quick Start

import { setProvider } from '@molecule/api-pdf'
import { provider } from '@molecule/api-pdf-puppeteer'

setProvider(provider)

Type

provider

Installation

npm install @molecule/api-pdf-puppeteer @molecule/api-pdf pdf-lib puppeteer

API

Interfaces

PuppeteerPDFConfig

Configuration options for the Puppeteer PDF provider.

interface PuppeteerPDFConfig {
  /**
   * Custom Puppeteer launch arguments passed to `puppeteer.launch()`.
   * Useful for running in Docker or CI environments.
   *
   * @example `['--no-sandbox', '--disable-setuid-sandbox']`
   */
  launchArgs?: string[]

  /**
   * Path to a custom Chromium/Chrome executable.
   * If omitted, Puppeteer uses its bundled browser.
   */
  executablePath?: string

  /**
   * Whether to run in headless mode. Defaults to `true`.
   */
  headless?: boolean

  /**
   * Timeout in milliseconds for page navigation and PDF generation.
   * Defaults to `30000` (30 seconds).
   */
  timeout?: number

  /**
   * Whether to reuse a single browser instance across calls.
   * Improves performance for batch operations. Defaults to `true`.
   */
  reuseBrowser?: boolean

  /**
   * Opening delimiter for simple template interpolation in `fromTemplate`.
   * Defaults to `'{{'`.
   */
  templateOpenDelimiter?: string

  /**
   * Closing delimiter for simple template interpolation in `fromTemplate`.
   * Defaults to `'}}'`.
   */
  templateCloseDelimiter?: string
}

Functions

createProvider(config)

Creates a Puppeteer-backed PDF provider.

function createProvider(config?: PuppeteerPDFConfig): PDFProvider
  • config — Optional provider configuration (launch args, executable path, etc.).

Returns: A PDFProvider backed by Puppeteer and pdf-lib.

Constants

provider

The provider implementation, lazily initialized with default config.

const provider: PDFProvider

Core Interface

Implements @molecule/api-pdf interface.

Bond Wiring

Setup function to register this provider with the core interface:

import { setProvider } from '@molecule/api-pdf'
import { provider } from '@molecule/api-pdf-puppeteer'

export function setupPdfPuppeteer(): void {
  setProvider(provider)
}

Injection Notes

Requirements

Peer dependencies:

  • @molecule/api-pdf ^1.0.1

Runtime Dependencies

  • @molecule/api-pdf
  • pdf-lib
  • puppeteer

E2E Tests

Integration checklist — drive the real UI (live preview, no mocks), adapt each item to this app's actual documents (invoice, report, receipt, contract) and check every box off one by one. A box you can't check is an integration bug to fix — not a skip:

  • Every document the app generates has a working Download/Export control that returns a REAL PDF — not an HTML error page or a JSON-stringified Buffer. Inspect the actual response: Content-Type is application/pdf and the body's first bytes are the %PDF magic (hex 25 50 44 46). Fetch the endpoint and check both — a body that starts with < or { is a failure dressed up as a download.
  • Opening the downloaded PDF shows the record's real values (names, line items, dates, totals) — not placeholder/template text or a blank page.
  • Edit a record and re-export: the new PDF reflects the changed values, and two different records produce two visibly different PDFs (not the same cached bytes for every id).
  • If the app shows page previews or reads document info, it feature-detects (getProvider().toImages / .getMetadata) or bonds a provider that supports them — both are OPTIONAL and THROW on bonds that lack them (e.g. PDFKit), so a preview built on an unsupporting bond errors at runtime, not compile time.
  • Styled output (CSS layout, backgrounds, web fonts) actually renders — which requires a browser-engine bond (Puppeteer). On PDFKit the same HTML collapses to a plain-text approximation; if the design matters, that's the wrong bond.
  • Export is authorized: a signed-in user cannot fetch another user's document by guessing or incrementing an id — the endpoint scopes every PDF to its owner (a guessed id returns 403/404, never someone else's invoice).