node-red-contrib-pdf2text

A Node-RED node that extracts the text of a PDF using unpdf (PDF.js under the hood, no native dependency).

Features

• Extract text from a PDF supplied as a Buffer, Uint8Array or ArrayBuffer.
• Limit the extraction to a page range (First page / Last page).
• Result metadata available on msg.pdf2text.
• Explicit, actionable errors for invalid input (never a silent empty output).
• Pure JavaScript — works on Windows, Linux and macOS with no build step.

Requirements

• Node.js 18 or newer
• Node-RED 3.0 or newer

Installation

From the Node-RED editor: Manage palette → Install → search pdf2text.

Or from the command line, inside your Node-RED user directory (usually ~/.node-red):

cd ~/.node-red
npm install node-red-contrib-pdf2text

Node: pdf2text

Category: parser

Input

| Property | Type | Description | |-----------|------------------------------------------|-------------| | payload | Buffer / Uint8Array / ArrayBuffer | The PDF as binary data. Configurable via PDF property. |

A string is not accepted. Reading a PDF as text (utf8 or base64) produces empty or corrupt output, so the node raises an explicit error telling you how to fix it. Configure the upstream node so it outputs binary data — for a file node set Output to "a single Buffer object". If you only have a base64 string, decode it first, e.g. with a function node: msg.payload = Buffer.from(msg.payload, "base64"); return msg;.

Output

| Property | Type | Description | |------------|----------|-------------| | payload | string | The extracted text. Configurable via Output property. | | pdf2text | object | Result metadata (fixed property, see below). |

msg.pdf2text contains:

| Field | Type | Description | |--------------|-----------|-------------| | success | boolean | true when extraction succeeded. | | totalPages | number | Total pages in the document. | | pages | number | Number of pages actually extracted (the requested range). | | fromPage | number | First extracted page (1-based). | | toPage | number | Last extracted page (1-based). | | elapsedMs | number | Processing time in milliseconds. |

Options

| Field | Description | |-------|-------------| | PDF property | Message property holding the PDF. Default msg.payload. Uses the standard Node-RED property picker. | | Output property | Message property where the text is written. Default msg.payload. | | First page | First page to extract (1-based, inclusive). Empty = from page 1. | | Last page | Last page to extract (1-based, inclusive). Empty = to the last page. |

The text of the selected pages is joined into a single string with a blank line between pages (\n\n).

Status

The node reports its result in the workspace:

• extracting... (blue) — processing.
• OK pages <last>/<total> (green) — success, e.g. OK pages 3/3.
• invalid input / failed (red) — an error was raised.

Errors

The node raises an error (and emits no message) when:

• The input is empty, or is not a Buffer/Uint8Array/ArrayBuffer (e.g. a string).
• The PDF cannot be parsed (corrupt, or not actually a PDF).
• First page is beyond the document, or First page > Last page.

A Last page beyond the document is clamped to the last page (a warning is logged).

Example flow

Import examples/extract-pdf-text.json via Menu → Import. It reads a PDF from disk into a Buffer and extracts the text.

Prefer a file node with Output: a single Buffer object over a function node — it avoids loading the whole file with fs.readFileSync.

Security notes

• Input validation: only binary input is accepted; strings are rejected.
• Prototype pollution: configured property paths are validated and reserved keys (__proto__, prototype, constructor) are refused.
• Resource handling: the PDF.js document is destroyed after each message to avoid memory leaks on large documents.

A PDF is untrusted input. PDF.js is generally robust, but a malformed PDF can consume CPU/memory; process untrusted files on a worker/service you can restart, and consider a file-size limit upstream.

Development

npm install
npm test                                          # smoke test + Node-RED harness
node test/nr-harness.js "C:/path/to/file.pdf"     # harness against a real PDF

To develop against a local Node-RED, install the folder into your user directory:

cd ~/.node-red
npm install "C:/path/to/node-red-contrib-pdf2text"

Contributing

Issues and pull requests are welcome. Please run npm test before opening a PR.

License

MIT