DocTable
Best OCR for Arabic Documents in 2026
For turning Arabic photos, scans and PDFs into an Excel or Word file with no setup, DocTable is our pick, and we make it, so weigh that. For developers, Azure Document Intelligence (Arabic printed and handwritten) and Google Cloud Vision (the cheapest at volume) are better and far cheaper per page. For free offline OCR, use Tesseract.
Your first 3 pages are free — scan and download, no card needed. After that, $3 buys 20 pages (one-time, no subscription).
Comparison table
| Tool | Price (entry plan) | Free tier? | Reads photos/scans or PDF only? | Arabic | API |
|---|---|---|---|---|---|
| DocTable | First 3 pages free; then $3 for 20 pages (one-time, no subscription) | Yes: 3 pages incl. download; first page needs no account | Photos, scans and PDFs (up to 500 pages per file) | Arabic and Persian, including mixed-script | Yes (API + MCP server) |
| Azure Document Intelligence (Read / Layout) (source source source) | Pay per page; paid rate not captured when we checked | Yes: F0 tier, 500 pages/month (first 2 pages of each file) | PDF and images (JPEG, PNG, BMP, TIFF, HEIF) | Arabic and Persian printed text; Arabic handwriting (v4) | Yes (REST + SDKs) |
| Google Cloud Vision (Text / Document Text Detection) (source source) | $1.50 per 1,000 pages after the free allowance | Yes: first 1,000 units per month | Images and PDFs (each PDF page counts as one image) | Arabic and Persian both listed | Yes |
| ABBYY FineReader PDF (source source) | $99/yr Standard, $165/yr Corporate (Windows); $69/yr Mac | 7-day trial, no payment required | Not stated on the pricing page | Arabic (Saudi Arabia) listed; Persian not in the list | Not published for the desktop product |
| Tesseract (source source) | Free (Apache 2.0) | Free and open source | Not checked | Arabic (ara) and Persian (fas) data files exist | Command-line tool and C/C++ API |
| Microsoft Excel: Data from Picture (source) | No separate price on the support page (Excel for Microsoft 365) | Not applicable: part of Excel | Picture file, clipboard image, or phone camera (mobile) | Not supported: Arabic and Persian are not among the listed character sets | No |
“Not published” or “not stated” means the vendor page we read does not say. Each tool links to the page the facts came from.
The tools, in order
1. DocTable
DocTable reads Arabic, Persian and mixed-script documents straight from photos, scans and PDFs, and rebuilds the rows and columns into an Excel or Word file. Your first 3 pages are free including the download, and the first page needs no account. After that $3 buys 20 pages, once, with no subscription. Median scan time is about 30 seconds per page. The honest limits: no custom models, no workflow builder, no SLA, and at $0.15 per page on the smallest pack it is far more expensive per page than the cloud APIs below if you can write code.
2. Azure Document Intelligence (Read / Layout)
Microsoft documents Arabic and Persian for printed text and, in the current v4 model, Arabic handwriting too. Persian handwriting is not in the list. Layout returns tables and structure, and the free tier gives 500 pages a month, though it processes only the first two pages of each file. This is the strongest option here for developers who need Arabic handwriting or enterprise controls; it is a cloud service you build against, not a drag-and-drop converter. The per-page price was blank in the copy of the pricing page we could read, so check it before budgeting.
Sources: source source source. Prices and features checked October 2026.
3. Google Cloud Vision (Text / Document Text Detection)
Google lists both Arabic and Persian as supported languages and charges $1.50 per 1,000 pages after 1,000 free units a month. At that price it is two orders of magnitude cheaper per page than any consumer converter, including ours. The catch is what you get back: text and coordinates for a developer to process, not a finished spreadsheet. If you have engineers and volume, Google is the better value; if you have a folder of photos and want an Excel file today, it is not.
Sources: source source. Prices and features checked October 2026.
4. ABBYY FineReader PDF
FineReader is a long-established desktop OCR suite, and for offline work on a Windows PC it is hard to beat: nothing is uploaded, and you edit the result in a full PDF editor. Its published OCR language list includes Arabic (Saudi Arabia) but not Persian. ABBYY's own support pages say Farsi is available in its server and SDK products rather than the desktop app. It is an annual subscription, so it suits people who scan every week more than people with one folder of documents.
Sources: source source. Prices and features checked October 2026.
5. Tesseract
Tesseract is free, open source (Apache 2.0) and recognises more than 100 languages out of the box, with Arabic and Persian data files available. If you can run a command line and write some code, it costs nothing and nothing leaves your machine. It outputs text, not a structured table, so rebuilding rows and columns is your job, and accuracy on photos depends heavily on how you pre-process them.
Sources: source source. Prices and features checked October 2026.
6. Microsoft Excel: Data from Picture
If you already pay for Microsoft 365 and your source is a clean, cropped, head-on picture of a table in English, German, French, Spanish and other listed Latin-script languages, Excel's built-in Data > From Picture is convenient and costs no extra. Microsoft's page lists the supported character sets, and Arabic and Persian are not among them. It handles one picture at a time and asks you to crop to the data.
Sources: source. Prices and features checked October 2026.
How we compared
We looked at tools that can read Arabic script at all, then checked each vendor's own pricing page and language list in October 2026. Where a vendor does not publish a fact, the table says so rather than guessing. We have not run a head-to-head accuracy test on the same documents, so we do not rank on accuracy; the order reflects fit for the job (finished Excel/Word file first, then developer APIs, then desktop and DIY). DocTable is our product, and it is listed first because the page is about the job it was built for.
Related pages
Temporary processing — files are not archived. Always verify OCR output before financial or legal use. Vendor names and prices belong to their owners and can change; check the linked page before you buy.