* * * GUIDES * * *

How to OCR a receipt in Python.

Receipts are the worst OCR input there is: thermal fade, crumples, photos taken at an angle. Classic engines (pytesseract) need a preprocessing pipeline; a vision-language model just reads them. Here's the whole thing in Python.

01 / INSTALL AND GET A KEY

One dependency. Keys are free — 100 pages/month, no card — at pennyocr.com/dashboard.

pip install pennyocr
export PENNYOCR_API_KEY=pk_live_...

02 / READ THE RECEIPT

Three lines. Works on JPEG photos, PNG screenshots and PDF scans alike.

from pennyocr import PennyOCR

result = PennyOCR().ocr("receipt.jpg", format="text")
print(result.text)

03 / WHAT YOU GET BACK

Verbatim text in reading order. From a coffee-shop receipt photo:

BLUE BOTTLE COFFEE
315 LINDEN ST, OAKLAND CA
1x CAPPUCCINO 5.50
2x LATTE 11.00
SUBTOTAL 26.50
TAX (8.75%) 2.32
TOTAL 28.82

04 / PARSE THE AMOUNTS

Totals are the point of receipt OCR. A regex over clean text is usually all you need:

import re

amounts = re.findall(r"(\d+\.\d{2})", result.text)
total = max(map(float, amounts))  # the total is almost always the largest
print(f"total: ${total}")

05 / BATCH A FOLDER

from pathlib import Path

client = PennyOCR()
for f in Path("receipts/").glob("*.jpg"):
    text = client.ocr(f, format="text").text
    print(f.name, "->", text.splitlines()[0])  # $0.00075 per receipt

Cost check: 1,000 receipts = $0.75, and your first 100 each month are free. If you need markdown with table structure instead, drop format="text".

GET A FREE KEY — 100 PAGES/MONTH FULL API DOCS