Key takeaways
- OCR text is the machine-readable text recognised from an image, often stored as an invisible layer in a searchable PDF.
- In Adobe Acrobat, OCR lives in the Scan & OCR tool, under Recognize Text; the free Reader cannot OCR.
- Tesseract reads images, not PDFs: convert PDF pages to images first, or use OCRmyPDF to add a text layer.
- List Tesseract's parameters with tesseract --print-parameters.
- For invoices, use the PDF's own text layer when it exists and fall back to OCR only for scanned pages.
OCR text is the text that optical character recognition extracts from an image of a page, making a scan searchable, copyable and editable. In Adobe Acrobat, OCR is in the Scan & OCR tool, where Recognize Text turns a scanned PDF into a searchable one; the free Acrobat Reader cannot run OCR. This page also explains Adobe Scan, using any OCR scanner, running Tesseract on PDFs and finding its parameters, the best options for handwriting, and how to build invoice OCR in a Chrome extension, drawing on the page-batched OCR in my OCR & Speech Workspace.
What is OCR text?
OCR text is the machine-readable text produced by optical character recognition from an image of text. In a searchable PDF, it is usually an invisible text layer placed under the page image, so the page looks the same but you can search, select and copy.
You can tell whether a PDF has OCR text by trying to select a sentence or search for a word: if nothing selects, the PDF is image-only. OCR text quality varies with scan quality; errors often appear in numbers, names and tables. Software that indexes documents, such as search engines and RAG chatbots, needs this text, so OCR is usually the first step in processing scanned archives.
Where is OCR in Adobe Acrobat?
In Adobe Acrobat, open the PDF, go to All tools and choose Scan & OCR, then select Recognize text, In this file. Acrobat runs OCR and adds a searchable text layer. Menu names vary slightly between versions.
You can also set the OCR language and choose whether to keep the original image or convert pages to editable text. For editing, the Edit PDF tool runs OCR automatically on scanned pages so you can change text directly. These features require a paid Acrobat plan; Adobe's online tools also offer some conversions. If you only need occasional OCR, free alternatives exist, covered below.
How do I OCR a PDF in Adobe Acrobat?
Open the scanned PDF in Acrobat, choose Scan & OCR, then Recognize text and In this file, pick the language and pages, and run it. Save the file, and it becomes a searchable PDF.
Also asked as: how to ocr a pdf in adobe acrobat
Tips for better results:
- Set the correct document language, including other languages if mixed.
- Use Enhance scanned file first if the scan is skewed or faint.
- Check a few pages by searching for known words.
- For editable output, export to Word after OCR and review formatting.
- For many files, use Acrobat's batch actions.
If text is still poor, the scan resolution may be too low; rescanning at a higher resolution helps more than any setting.
Can you OCR in Adobe Reader?
No. The free Adobe Acrobat Reader can view, comment on and sign PDFs, but it cannot run OCR. OCR needs Acrobat Standard or Pro, or Adobe's paid online services.
Free alternatives: open the PDF from Google Drive with Google Docs, which runs OCR; use Microsoft Word to open and convert a scanned PDF; use OCRmyPDF, a free open-source command-line tool that adds a text layer; or use the Adobe Scan mobile app for documents you are scanning yourself. For sensitive documents, prefer offline tools over uploading to unknown websites.
What does OCR mean in Adobe Scan?
In Adobe Scan, Adobe's mobile scanning app, OCR means the app automatically recognises text in the documents you photograph, so the PDFs it creates are searchable and you can copy text from them.
Adobe Scan detects page edges, corrects perspective, enhances contrast and runs OCR on each page. Results sync to Adobe's cloud, where you can open them in Acrobat. Good capture habits improve OCR: flat pages, even light without shadows, the phone parallel to the page, and one page per shot. Check the app's settings for OCR language. As with any cloud service, avoid scanning highly sensitive documents unless your organisation approves it.
How do I use an OCR text scanner?
Open the scanner app, capture the page in good light, let it crop and straighten the image, confirm OCR is on and the language is correct, then save as a searchable PDF or copy the recognised text.
Also asked as: how to use ocr text scanner
Steps that work in most apps, such as Adobe Scan, Google Lens, Microsoft OneNote or your phone's notes app:
- Place the page on a contrasting, flat surface.
- Hold the phone directly above, parallel to the page.
- Capture each page separately; review crops.
- Choose PDF for documents, text for quick copying.
- Search or select text to confirm OCR worked.
For handwriting, accuracy drops, so check carefully. For multi-page documents, scan into one PDF and name it clearly.
Does Tesseract OCR convert a PDF to text?
Not directly. Tesseract reads images, so you first convert PDF pages to images, then run Tesseract on each. Alternatively, OCRmyPDF uses Tesseract to add a text layer to the whole PDF in one command.
Two approaches:
ocrmypdf --language eng scanned.pdf searchable.pdf # searchable PDF in one step
from pdf2image import convert_from_path
import pytesseract
pages = convert_from_path("scanned.pdf", dpi=300)
text = "\n".join(pytesseract.image_to_string(p, lang="eng") for p in pages)
If the PDF already contains text, as most digitally created PDFs do, extract it directly with a PDF library instead, which is faster and exact.
How do I find the parameters supported in Tesseract's config file?
Run tesseract --print-parameters to list every parameter with its default value and a short description. Then set values in a config file or pass them on the command line with -c name=value.
Also asked as: how to find parameters supported in tesseract ocr config file
Useful options to know: --psm sets page segmentation mode, for example 6 for a single block of text or 7 for a single line; --oem selects the OCR engine mode; -c tessedit_char_whitelist=0123456789 restricts recognised characters, useful for digits; and preserve_interword_spaces=1 keeps spacing. Config files placed in Tesseract's configs folder can be referenced by name. Change one setting at a time and measure accuracy on a small sample.
What is the best OCR model for handwritten text?
Modern vision-language models and specialised handwriting recognition services handle handwriting best. Traditional engines such as Tesseract are designed for print and perform poorly on handwriting.
Options include large multimodal models from major AI providers, cloud document services with handwriting support, and open models such as TrOCR for line-level recognition. Results depend heavily on writing neatness, language and scan quality. Always review handwritten OCR, ask models to mark uncertain words, and keep the original image for checking. For notes taken with a stylus, built-in conversion in notes apps is often more accurate than OCR of photos, because it uses pen strokes.
How do I OCR invoices in a Chrome extension?
First try the PDF's own text layer, using a PDF library such as PDF.js in the browser; only if a page has no text, send its image to an OCR service on your server. Then parse fields such as invoice number, date and total with rules or a model, and validate them.
Also asked as: how to ocr invoices in a chrome extension: pdf text layer, server ocr, and field parsing
A robust design:
- Text layer first: extract text and positions with PDF.js; most digital invoices have it.
- Server OCR fallback: send scanned pages to a backend running Tesseract or a document AI service.
- Field parsing: regular expressions for dates and amounts, or an LLM with a strict JSON schema.
- Validation: totals equal the sum of line items, dates are valid, required fields present.
- Privacy: tell users what leaves the browser, and process locally where possible.
That is the pattern behind DocuLens AI: extraction with schema validation before anything is trusted.
Sources
GenAI engineer and AI product builder with 2+ years shipping production-grade AI systems: RAG pipelines, fine-tuned LLMs, hybrid retrieval and multi-modal apps across vision, speech and OCR, architected end to end from ingestion to deployment. Leads engineering for SCULT INDIA's 14-member team, founded the 500+ member TechVerse Enclave community and has mentored 200+ students. Three hackathon first prizes: Changethon 2025 (IIT Roorkee), Product Genesis at Vividhotsava 2025 (CSJMU) and BYTEBATTLE (MeetKats).
Open to GenAI roles, hackathon judging, mentorship sessions and guest talks at colleges. For campus sessions the inviting institution covers the visit end to end: travel (a First AC train ticket for journeys up to five hours, a flight for anything longer), accommodation, meals, and pickup and drop; the honorarium and workshop fees are negotiable on WhatsApp or on the call, and requests are welcome from institutions that can cover the visit this way. Email pranjulrathour41@gmail.com.
