FormatForge logoFormatForge

Procurement Document Tools

Multilingual OCR for English, Greek, Hindi, Arabic and Europe

Turn scanned procurement documents into editable text before comparing purchase orders, delivery notes and supplier invoices. Processing runs in the browser and supports nine languages, including mixed-language documents with up to three selected language models.

Browser-based multilingual OCR

Upload a business document

Extract searchable text from invoices, purchase orders, delivery notes, receipts and scanned forms in nine supported languages.

OCR settings

Fast mode

Select only the languages actually present. Up to three languages can be combined.

Performance tip: For faster and more accurate OCR, select only the languages present in your document. A single language usually gives the best speed and accuracy.
Accuracy and speed: one language is fastest. Use two or three only for genuinely mixed documents. Clear, straight scans at 200–300 DPI work best.
OCR accuracy notice: OCR results may contain errors. Verify quantities, prices, tax values, supplier details and totals before business, accounting or legal use.

Private, client-side document processing

Your selected PDF or image is processed inside your browser. FormatForge does not upload the document to its application server, create a document account, or save the extracted text. The OCR engine and selected language model are downloaded from their public delivery networks, while your document stays in the current browser session.

✓ No document upload to FormatForge

✓ No document or OCR-result storage

✓ No registration required

✓ Clear the page or close the tab to discard session data

For highly regulated or classified documents, follow your organisation's security policy and approved software requirements before processing any file.

Built for real procurement documents

Supplier documents frequently arrive as scanned PDFs or phone photographs rather than structured spreadsheets. This OCR tool gives warehouse, procurement and accounts-payable teams a practical first step: extract the visible text, inspect uncertain values and prepare the document for matching or data entry.

English documents

Purchase orders, invoices, packing lists, delivery notes, item descriptions and general forms.

Greek — Ελληνικά

Greek supplier invoices, δελτία αποστολής, quantities, VAT labels and local product descriptions.

Hindi — हिन्दी

Hindi invoices, stock records, purchase documents and mixed Hindi–English descriptions.

Arabic — العربية

Arabic and bilingual documents from the UAE, Bahrain, Saudi Arabia, Qatar, Oman and Kuwait.

German, French and Dutch

Procurement documents used across Germany, France, Austria, Belgium and the Netherlands.

Spanish and Italian

Invoices, delivery notes and supplier documents from Spain, Italy and international partners.

Recommended workflow

  1. Select one language for the fastest result, or combine up to three for mixed documents.
  2. Upload a straight, readable scan or image.
  3. Extract the text and review confidence.
  4. Check item codes, quantities, prices, VAT and decimal separators.
  5. Copy or download the corrected text for comparison or ERP entry.

What requires manual verification?

  • • Similar characters such as O/0, I/1 and Greek/Latin look-alikes
  • • Comma and period decimal separators
  • • Product codes, barcodes and long invoice numbers
  • • Rotated, faint, stamped or handwritten text
  • • Table columns that visually overlap
  • • Totals, VAT values and payment-sensitive fields

Next step: compare procurement documents

OCR extracts the content. The next stage is sequence-independent matching, where products are aligned using barcode, SKU, supplier code, normalised description and quantity rather than row position.

Frequently asked questions

Does the OCR support Greek invoices and delivery notes?

Yes. Choose Greek or English + Greek before processing. The tool can recognise Greek characters, numbers, product descriptions and common business-document text, although every result should be checked against the source scan.

Can it read Hindi documents?

Yes. Select Hindi, or combine Hindi with English, for documents containing Devanagari and Latin text. Clear scans and a small language selection improve recognition.

Does it support Arabic invoices from Dubai, Bahrain and the Gulf?

Yes. Arabic OCR supports right-to-left text used across the UAE, Bahrain, Saudi Arabia, Qatar, Oman and Kuwait. Select Arabic alone for speed or Arabic with English for bilingual documents.

Which European languages are included?

The tool supports German, French, Spanish, Italian and Dutch in addition to English and Greek. You can combine up to three languages when supplier documents are genuinely multilingual.

Are my documents uploaded to FormatForge?

The document is processed in your browser. The OCR engine and selected language models are downloaded from public delivery networks when required, but FormatForge does not send the document to its own server.

Can I use the extracted text directly for invoice approval?

No. OCR can confuse similar characters, decimal separators, product codes and table columns. Verify all important values against the original document before receiving goods, posting an invoice or making a payment.

Why is table formatting not perfectly preserved?

OCR recognises characters rather than reconstructing every visual table cell. The extracted text is intended for search, copying and preparation for the procurement document-matching workflow.

What files are supported?

You can process PDF, PNG, JPG, JPEG, WebP and BMP files up to 25 MB. For long PDFs, select a practical page limit to control processing time and browser memory.