How to Extract Data From a PDF

Upload a PDF, create an extraction job, wait for completion, and query the structured rows.

Use this procedure after you create a table.

You need:

a project API key in PDFPARSE_API_KEY
the table slug from the create-table response
a synthetic or redacted PDF
Node.js 20 or later

1. Install the upload client

PdfParse uses the TUS resumable upload protocol.

npm install tus-js-client

2. Upload the PDF

Save this file as extract-pdf.mjs:

import fs from "node:fs";
import path from "node:path";
import { Upload } from "tus-js-client";

const apiKey = process.env.PDFPARSE_API_KEY;
const tableSlug = process.env.PDFPARSE_TABLE_SLUG;
const filePath = process.argv[2];

if (!apiKey || !tableSlug || !filePath) {
  throw new Error(
    "Set PDFPARSE_API_KEY and PDFPARSE_TABLE_SLUG, then provide a PDF path.",
  );
}

const file = fs.createReadStream(filePath);
const size = fs.statSync(filePath).size;
const filename = path.basename(filePath);

const uploadResult = await new Promise((resolve, reject) => {
  const upload = new Upload(file, {
    endpoint: "https://api.pdfparse.net/v1/uploads",
    uploadSize: size,
    headers: {
      Authorization: `Bearer ${apiKey}`,
    },
    metadata: {
      filename,
      name: filename,
      type: "application/pdf",
      file_size: String(size),
      tableSlug,
    },
    retryDelays: [0, 1000, 3000, 5000],
    removeFingerprintOnSuccess: true,
    onError: reject,
    onSuccess() {
      const encodedUploadId = upload.url?.split("?")[0]?.split("/").filter(Boolean).at(-1);
      if (!encodedUploadId) {
        reject(new Error("The upload did not return an upload ID."));
        return;
      }
      const uploadId = decodeURIComponent(encodedUploadId);

      resolve({
        uploadId,
        documentKey: uploadId,
      });
    },
  });

  upload.start();
});

3. Create the extraction job

Add this code after the upload code:

const baseUrl = "https://api.pdfparse.net/v1";
const headers = {
  Authorization: `Bearer ${apiKey}`,
  "Content-Type": "application/json",
};

const jobResponse = await fetch(`${baseUrl}/jobs`, {
  method: "POST",
  headers: {
    ...headers,
    "Idempotency-Key": crypto.randomUUID(),
  },
  body: JSON.stringify({
    tableSlug,
    documentKeys: [uploadResult.documentKey],
  }),
});

if (!jobResponse.ok) {
  throw new Error(await jobResponse.text());
}

const job = await jobResponse.json();
console.log(`Created job ${job.id}`);

The job runs asynchronously.

4. Wait for completion

Add a bounded polling loop for a basic integration:

const wait = (milliseconds) =>
  new Promise((resolve) => setTimeout(resolve, milliseconds));

let completedJob;

for (let attempt = 0; attempt < 60; attempt += 1) {
  const response = await fetch(`${baseUrl}/jobs/${job.id}`, { headers });

  if (!response.ok) {
    throw new Error(await response.text());
  }

  const currentJob = await response.json();

  if (currentJob.status === "completed") {
    completedJob = currentJob;
    break;
  }

  if (currentJob.status === "failed") {
    throw new Error(`Job ${job.id} failed.`);
  }

  await wait(2000);
}

if (!completedJob) {
  throw new Error(`Job ${job.id} did not finish in time.`);
}

Use a webhook instead of polling for a production integration.

5. Query the extracted rows

The completed job response contains table_id.

const rowsResponse = await fetch(`${baseUrl}/rows/search`, {
  method: "POST",
  headers,
  body: JSON.stringify({
    table_id: completedJob.table_id,
    filters: [],
    page: 1,
    pageSize: 50,
  }),
});

if (!rowsResponse.ok) {
  throw new Error(await rowsResponse.text());
}

const rows = await rowsResponse.json();
console.log(JSON.stringify(rows, null, 2));

Run the complete script:

export PDFPARSE_TABLE_SLUG="invoices-replace_with_your_slug"
node extract-pdf.mjs ./sample-invoice.pdf

6. Validate the result

Check the extracted values against the source PDF.

Confirm that:

required fields are present
codes keep leading zeros
dates use the expected format
amounts use the expected decimal value
repeating rows have the expected count

Do not send unvalidated OCR data directly to a financial or operational system.

For production event delivery, use the webhook endpoints in the API reference. For authentication errors, read Authentication.