Image OCR Text Extractor — Images & Scanned PDFs avatar

Image OCR Text Extractor — Images & Scanned PDFs

Pricing

from $33.50 / 1,000 ocr pages

Go to Apify Store
Image OCR Text Extractor — Images & Scanned PDFs

Image OCR Text Extractor — Images & Scanned PDFs

OCR images and scanned PDFs by link with Tesseract: JPEG, PNG, WebP, TIFF and PDF pages, 24 languages. Text line by line with word count, confidence and optional word positions. Pages with no text are not billed. Part of NexGen Signal — data tools and official-source data products

Pricing

from $33.50 / 1,000 ocr pages

Rating

0.0

(0)

Developer

NexGen Signal

NexGen Signal

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

a month ago

Last modified

Features

Limited permissions

Share

🔎 Image OCR Text Extractor — Images & Scanned PDFs

Give image links (JPEG, PNG, GIF, WebP, BMP, TIFF) or PDF links, including scanned PDFs, and get the text in each image or PDF page as one row: text line by line with paragraphs, word and line counts, average confidence, and optionally every word with its confidence and position.

OCR runs inside the actor with Tesseract 5 and 24 language packs (English, German, French, Spanish, Italian, Portuguese, Dutch, Polish, Swedish, Danish, Norwegian, Finnish, Czech, Turkish, Russian, Ukrainian, Arabic, Hindi, Chinese simplified and traditional, Japanese, Korean, Vietnamese, Indonesian); pick up to five per run. Images are converted to grayscale and small images are enlarged before reading; PDF pages are rendered at 300 DPI by default.

The unit of output is one image or PDF page delivered. Pages where no text is found are not billed.

Who buys this and for what job

  • Back-office and finance teams reading scanned letters, forms and receipts they hold the rights to.
  • Archivists and researchers making historical scans searchable.
  • AI builders adding scanned documents to a knowledge base.
  • Developers needing an OCR step without running Tesseract themselves.

Buyers usually search for this as "OCR API", "image to text" or "scanned PDF to text".

📊 Sample Output

Table of OCR results: link, page number, word count, mean confidence and the start of the text.

source_typepage_numberword_countmean_confidencetext
image153690.3Albert Einstein Old Grove Ra. Wassau Point Peconic, Long Is …
pdf_page344294.3Form W-4 (2026) Step 2(b)— Multiple Jobs Worksheet (Keep fo …
pdf_page539989.3Form W-4 (2026) Higher Paying Job Annual Taxable Wage & Sal …
pdf_page181792.9Form W-4 Department of the Treasury Internal Revenue Servic …
pdf_page496392.9Form W-4 (2026) Step 4(b)— Deductions Worksheet (Keep for y …

Rows from live run ntbOdvzs0bJP180ct on Apify (the Store example: the 1939 Einstein–Szilard letter (a typed scan) and the five pages of the IRS W-4 form). Notice the confidence: about 90 on the 1939 typewritten scan and 89 to 96 on the printed form pages.

✅ What you get

Every field of each record:

  • record_id — file hash plus page number.
  • url, final_url, found_on_page — the link, where it was served from, and the page it was found on (with followImageLinks).
  • source_type, page_number, page_count — image or pdf_page, and its position.
  • width, height — pixels (PDF pages as rendered).
  • languages — OCR languages used.
  • text, text_chars, word_count, line_count — the text and its size.
  • mean_confidence — average word confidence, 0 to 100.
  • words — each word with confidence and bounding box (with includeWords).
  • sha256, processed_at — file hash and when it was read (UTC).

⚙️ Sample inputs

A scanned letter and a PDF (the Store example)

{
"imageUrls": ["https://upload.wikimedia.org/wikipedia/commons/b/bf/Einstein-Roosevelt-letter.png", "https://www.irs.gov/pub/irs-pdf/fw4.pdf"],
"maxPages": 10
}

One image and five PDF pages.

Word positions in two languages

{
"imageUrls": ["https://upload.wikimedia.org/wikipedia/commons/b/bf/Einstein-Roosevelt-letter.png"],
"languages": ["eng", "fra"],
"includeWords": true,
"maxPages": 3
}

Each word with its box, for highlighting or redaction tools.

Labels and signs

{
"imageUrls": ["https://example.com/your-image.png"],
"pageSegmentation": "sparse",
"maxPages": 10
}

Sparse mode finds scattered text such as labels and screenshots.

Input fields:

FieldMeaning
imageUrlsimage or PDF links, or pages with followImageLinks (up to 5,000).
languages1 to 5 language codes (default eng).
pageSegmentationauto, single_block, single_column, sparse, single_line.
includeWordsadd word boxes and confidences (default false).
followImageLinksread images and PDFs on pages you give (default false).
maxPdfPagespages read per PDF (default 50).
pdfDpiPDF render resolution, 72 to 400 (default 300).
maxFileMbskip files bigger than this (default 50).
maxPagesstop after this many delivered pages (default 1,000). Each delivered page is billed.

🧾 JSON sample record

{
"record_id": "9958a33916fa3ea43d2c-1",
"url": "https://upload.wikimedia.org/wikipedia/commons/b/bf/Einstein-Roosevelt-letter.png",
"final_url": "https://upload.wikimedia.org/wikipedia/commons/b/bf/Einstein-Roosevelt-letter.png",
"found_on_page": null,
"source_type": "image",
"page_number": 1,
"page_count": 1,
"width": 1500,
"height": 881,
"languages": [
"eng",
"fra"
],
"text": "Albert Einstein\n\nOld Grove Rd.\nHassau Point\nPeconic, Long Island\n\nAugust 2nd, 1939\n\nF.D. Roosevelt,\n\nPresident of the United States,\nWaite Hous:\n\nington, D.C.\n\nSirs\n\nSome recent work by E.Permi and L. Sailard, which has been com-\nmunicated to me in manuscript, leads me to expect that the element uran-\nium may be turned into a new and important source of energy in the in-\nmediate future. Certain aspects of the situation which has arisen seem\nto call for watchfulness and, if necessary, quick actio …",
"text_chars": 3180,
"word_count": 536,
"line_count": 61,
"mean_confidence": 92,
"words": [
{
"text": "Albert",
"confidence": 95.3,
"left": 465,
"top": 42,
"width": 52,
"height": 11,
"line": 1
},
{
"text": "Einstein",
"confidence": 74.2,
"left": 527,
"top": 42,
"width": 70,
"height": 11,
"line": 1
},
{
"text": "Old",
"confidence": 87.9,
"left": 466,
"top": 56,
"width": 26,
"height": 11,
"line": 2
},
"… more words"
],
"sha256": "9958a33916fa3ea43d2cb646c1e40d1c1cdf96b411d0e8d90efc3da32db8619f",
"processed_at": "2026-10-17T04:11:23Z"
}

🔧 How it works

Files are downloaded 6 at a time; pages are read in parallel up to the run's CPU count. Tesseract's word table (TSV) is turned into lines and paragraphs; the average of the word confidences is reported.

Every run reads each host's robots.txt first: a path it disallows is skipped, listed in the run receipt and never charged. Only public internet addresses are fetched (no private, loopback or cloud-metadata addresses, checked again on every redirect). No logins, no CAPTCHA solving, no proxies.

Delivery is pushed first and charged after. The run receipt (RUN_RECEIPT) lists every link with its outcome and pages delivered, page outcomes with counts (delivered, no text found, unreadable), the robots.txt result per host, and the delivered and charged counts.

💰 Pricing example

Per page delivered, no start fee, no monthly charge:

Apify planPrice per page
Free$0.05
Bronze$0.045
Silver$0.04
Gold, Platinum, Diamond$0.0335
  • 10 pages: $0.50 on the Free plan, $0.335 on Gold.
  • 1,000 pages: $50.00 on Free, $33.50 on Gold.
  • 10,000 pages: $500.00 on Free, $335.00 on Gold.

Coverage and size. Proof runs on 17 Oct 2026: 40 pages of IRS Publication 17 at 200 DPI in about 4 minutes on the default 4 GB memory (more memory gives more CPU and runs faster); the Store example (6 pages) in about 20 seconds. Handwriting is not read reliably.

Images belong to their owners. You choose the links. Make sure you have the right to download and use the content you point it at; the actor fetches only what you give it (and, where you turn it on, what those pages link to).

Every run reads each host's robots.txt first: a path it disallows is skipped, listed in the run receipt and never charged. Only public internet addresses are fetched (no private, loopback or cloud-metadata addresses, checked again on every redirect). No logins, no CAPTCHA solving, no proxies.

❓ FAQ

Does it read handwriting? Not reliably. Tesseract is built for printed and typed text.

Why is a page missing? Pages where OCR found no text are not delivered or billed; RUN_RECEIPT counts them.

Can I make it faster? Give the run more memory (Apify gives more CPU with more memory) or lower pdfDpi.

🆘 Troubleshooting

  • Low confidence or garbled text — Check the languages; try pageSegmentation single_block or sparse; raise pdfDpi for small print.
  • not an image or PDF — The link returned something else, such as a web page; use followImageLinks for pages.
  • SOURCE UNAVAILABLE — None of your links could be reached after retries. Nothing was charged.

🏢 About NexGen Signal — data tools and official-source data products on Apify. Browse the catalog: apify.com/nexgensignal

⭐ Found this useful?

If this actor saved you time, a rating on its Apify Store page helps other buyers find it.

Part of NexGen Signal — data tools and official-source data products.