Joomla OCR: text recognition for receipts, PDFs and photos
J!DMS brings OCR text recognition to Joomla 6. Scanned invoices, till receipts, photos and PDFs are read automatically during upload and can then be searched in full text – without extra software on the server and without a paid OCR service.
Can Joomla recognise text out of the box?
No. Joomla stores uploaded images and PDFs as files but does not read their content. A scanned receipt is therefore invisible to the Joomla search. J!DMS closes this gap: it recognises the text of every document and stores it together with the file in your Joomla database.
How text recognition works in J!DMS
- PDFs with a text layer – for PDFs from your accounting software, online shop or office program, J!DMS reads the existing text directly. That is fast and error-free.
- Scanned PDFs and photos – if a PDF page has no text, it is rendered as an image and recognised with OCR. The same applies to JPG, PNG, WebP and GIF, for example a phone photo of a till receipt. J!DMS checks page by page, so mixed PDFs are no problem.
- Store and find – the file, the recognised text and a thumbnail are stored together in the database. You can then find the receipt by any word in it: supplier, invoice number, amount or item.
Why recognition runs in the browser
Classic OCR solutions need a program such as Tesseract on the server. Most web hosting plans don't allow that. J!DMS therefore uses Tesseract.js and pdf.js directly in the browser of the person uploading the receipts. This has three advantages:
- It works on any ordinary Joomla hosting, including shared hosting.
- The server is not loaded, even with many documents.
- There is no external OCR service, no API key and no cost per page.
If you can install Tesseract and Poppler on your own server, you can additionally enable server recognition as a fallback in the options.
Languages, page count and settings
- By default J!DMS recognises German and English at the same time. You can add further Tesseract languages in the options.
- For long PDFs up to 50 pages are read by default, configurable up to 500.
- The recognised text can be corrected in the edit view and re-recognised at any time – for one or many receipts at once.
Privacy
During text recognition the documents never leave your browser and your Joomla installation. There is no cloud the receipts are sent to. Only the program libraries (Tesseract.js, pdf.js and the language data) are loaded from a public CDN by default. For strictly GDPR-compliant operation, store these files on your own server and enter their addresses in the options.
Where text recognition reaches its limits
Printed text on clean scans (from about 300 dpi) and straight photos of receipts are recognised reliably. Handwriting, heavily faded thermal receipts, crumpled receipts and very small print are difficult. Even then the receipt can still be found by title, note, category and receipt date, and you can add important details to the text by hand.
Frequently asked questions
Does text recognition cost extra?
No. OCR is included in J!DMS Free (up to 20 documents) and J!DMS Pro; there are no per-page fees.
Which files are recognised?
PDF (with and without a text layer), JPG, PNG, WebP and GIF.
Do I need Tesseract on the server?
No. Recognition runs in the browser. Server recognition is only an optional fallback.
What about receipts that are already in the archive?
Select them in the list and start text recognition again. J!DMS then replaces text and thumbnail.
Which Joomla version do I need?
Joomla 6 with PHP 8.3 or later and MySQL or MariaDB.
Try it now
The free version includes all features including text recognition for up to 20 documents.