PDF to Text Python Examples

Complete Python examples for common PDF to Text tasks. See the Python SDK guide for installation and result handling.

You can run the examples below with the displayed API credentials. Update the input URLs and filenames as needed. Each Python example can be copied and run independently.

Basic examples

Choose file, in-memory, or output-stream handling for the text result. Stream examples use a binary file stream as their destination.

The memory variants return UTF-8 bytes. Decode them when your application needs a text string.

Convert a PDF file

Upload an existing invoice.pdf file.

Save to a file

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

client.convertFileToFile('invoice.pdf', 'invoice.txt')

Keep in memory

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

text_bytes = client.convertFile('invoice.pdf')  # bytes
print(len(text_bytes))

Write to a stream

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

with open('invoice.txt', 'wb') as output_stream:
    client.convertFileToStream('invoice.pdf', output_stream)

Convert a URL

Fetch a PDF from its URL.

Save to a file

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

client.convertUrlToFile('https://your-server.com/invoice.pdf', 'invoice.txt')

Keep in memory

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

text_bytes = client.convertUrl('https://your-server.com/invoice.pdf')  # bytes
print(len(text_bytes))

Write to a stream

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

with open('invoice.txt', 'wb') as output_stream:
    client.convertUrlToStream(
        'https://your-server.com/invoice.pdf', output_stream
    )

Convert raw PDF bytes

Read invoice.pdf as binary data. In an application, pass the bytes you already hold in memory instead.

Save to a file

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

with open('invoice.pdf', 'rb') as source_file:
    pdf_data_1 = source_file.read()  # bytes
client.convertRawDataToFile(pdf_data_1, 'invoice.txt')

Keep in memory

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

with open('invoice.pdf', 'rb') as source_file:
    pdf_data_1 = source_file.read()  # bytes
text_bytes = client.convertRawData(pdf_data_1)  # bytes
print(len(text_bytes))

Write to a stream

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

with open('invoice.pdf', 'rb') as source_file:
    pdf_data_1 = source_file.read()  # bytes
with open('invoice.txt', 'wb') as output_stream:
    client.convertRawDataToStream(pdf_data_1, output_stream)

Read from an input stream

Open invoice.pdf as a binary input stream and save the result to invoice.txt.

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

with open('invoice.pdf', 'rb') as input_stream:
    client.convertStreamToFile(input_stream, 'invoice.txt')

Decode extracted text

Convert a PDF URL, decode the returned bytes as UTF-8, and print the resulting string.

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

text_bytes = client.convertUrl('https://your-server.com/invoice.pdf')  # bytes
text = text_bytes.decode('utf-8')
print(text)

Customize text extraction

These examples upload extraction.pdf, a three-page document with a two-column article, wrapped paragraphs, and invoice fields.

Extract selected pages

Extract pages 1 and 3 of extraction.pdf into excerpt.txt.

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

client.setPrintPageRange('1,3')
client.convertFileToFile('extraction.pdf', 'excerpt.txt')

Extract text without layout spacing

Extract the article on page 1 into article.txt without spacing that recreates the PDF's columns. For tables, the default layout mode usually preserves column alignment better.

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

client.setNoLayout(True)
client.setPrintPageRange('1')
client.convertFileToFile('extraction.pdf', 'article.txt')

Reconstruct wrapped paragraphs

Extract page 2 into paragraphs.txt, joining wrapped lines with paragraph detection and removing line-end hyphens. For example, infor- followed by mation becomes information. Paragraph detection depends on the document's layout.

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

client.setParagraphMode('bounding-box')
client.setPrintPageRange('2')
client.setRemoveHyphenation(True)
client.convertFileToFile('extraction.pdf', 'paragraphs.txt')

Extract text from a page region

Crop away the headers and footers on every page, or extract only the customer, invoice, and reference fields on page 3.

The coordinates below match a 612 × 792 point (US Letter) page. Measure X and Y from the top-left corner; 72 points equal one inch. Adjust the regions for your PDF.

Exclude headers and footers

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

client.setCropAreaHeight(648)
client.setCropAreaWidth(540)
client.setCropAreaX(36)
client.setCropAreaY(72)
client.convertFileToFile('extraction.pdf', 'body.txt')

Extract an invoice field block

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

client.setCropAreaHeight(70)
client.setCropAreaWidth(300)
client.setCropAreaX(36)
client.setCropAreaY(290)
client.setPrintPageRange('3')
client.convertFileToFile('extraction.pdf', 'fields.txt')

Keep page boundaries

The default output has no page separators. Insert form-feed characters (\f) for processing text by page, or add a visible --- PAGE BREAK --- marker on its own line.

Use form-feed separators

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

client.setPageBreakMode('default')
client.convertFileToFile('extraction.pdf', 'pages.txt')

Use visible separators

import pdfcrowd

client = pdfcrowd.PdfToTextClient('demo', 'demo')

client.setCustomPageBreak("""
--- PAGE BREAK ---
""")
client.setPageBreakMode('custom')
client.convertFileToFile('extraction.pdf', 'separated.txt')