Inkspan › Guides › How to extract text from a PDF in Node.js

How to extract text from a PDF in Node.js

The /v1/extract/text endpoint returns the text content of a PDF as JSON, split per page. Each entry carries its page number and the text found on it, which keeps positional context you would lose by concatenating the whole document into one string.

This reads the text layer rather than performing OCR. Born-digital PDFs — anything produced by software, which covers most invoices, statements, and exports — extract cleanly and quickly.

Also in: cURL · Python

Node.js example

import fs from 'node:fs';

const fd = new FormData();
fd.append('file', new Blob([fs.readFileSync('input.pdf')]), 'input.pdf');

const res = await fetch('https://api.inkspan.dev/v1/extract/text', {
  method: 'POST',
  headers: { 'X-API-Key': process.env.INKSPAN_KEY },
  body: fd,
});
if (!res.ok) throw new Error(await res.text());
console.log(await res.json());

Request parameters

Name	Type	Required	Description
`file`	file	yes	The PDF to read text from.

Response

Returns application/json:

{
  "page_count": 2,
  "pages": [
    {
      "page": 1,
      "text": "Invoice INV-1042\nTotal: $61.00"
    },
    {
      "page": 2,
      "text": "Terms and conditions\u2026"
    }
  ]
}

Things that catch people out

This is not OCR. A scanned page with no text layer returns an empty string for that page rather than an error.
Check for empty text before trusting the result — it is the signal that a document is scanned and needs a different pipeline.
Reading order follows the PDF's internal content order, which for multi-column layouts may not match visual order.
The response is JSON, so parse it rather than writing the body to a file.

Errors

Status	Meaning	Fix
`401`	Invalid or missing API key	Send your key in the `X-API-Key` header.
`402`	Monthly quota exceeded	The response reports used/quota. Upgrade at inkspan.dev/#pricing.
`422`	Missing or invalid parameter	Check required fields in the table above — the body names the offending field.

Failed requests do not count against your quota; only successful ones are metered.

Frequently asked questions

Does it handle scanned PDFs?

No. Extraction reads the embedded text layer; a scan without one returns empty pages. Rasterise with /v1/pdf/to-images and run OCR or a vision model if you need that.

How do I tell a scanned PDF from a digital one?

Extract and check whether the text is empty. Empty pages with a non-trivial page count almost always mean a scan.

Is layout preserved?

Text comes back per page in the document's internal order. For tabular data use /v1/extract/tables, which reconstructs row and column structure.

Get a free API key →

Related Extract guides: extract tables from a PDF