Can ATS Systems Actually Read an Arabic CV?

Arabic CVs do get read by applicant tracking systems. The script is not the problem, and the widespread belief that ATS software “cannot handle Arabic” sends people into the wrong fix — submitting an English CV to an employer whose entire operation runs in Arabic.

The real failures are mechanical and specific: fonts that were never embedded, glyph shapes stored instead of letters, layouts read in the wrong direction, and digits that belong to a different character set. Every one of them is visible in under a minute, and every one is fixable without changing a word of your content.

Quick answer

• Arabic text parses fine in the systems large Saudi employers use. Right-to-left direction is handled.
• What breaks it: unembedded or decorative fonts, design-tool exports, tables, two-column layouts, and scanned PDFs.
• Arabic-Indic digits (٢٠٢٤) do not match a search for 2024. Use 0-9 everywhere.
• Never use side-by-side Arabic/English columns. It is the worst possible format.
• The test: open the PDF, select all, copy, paste into a plain text file. What you see is what the system sees.

For the underlying build rules that apply in both languages, read how to write a CV that gets past the ATS, and for the language decision itself, Arabic or English.

What actually happens when a parser opens your PDF

A PDF is not a document in the way a Word file is. It is a set of drawing instructions: put this glyph at this coordinate, in this font, at this size. A parser reconstructs text by collecting those glyphs and guessing the reading order from their positions.

For Latin text that guess is usually easy. For Arabic, four extra things can go wrong:

1. Shaping. Arabic letters change form depending on position — initial, medial, final, isolated. Many exporters store the shaped presentation form of each letter instead of the base letter. Extraction then produces text that looks Arabic but does not match a search for the normal word.
2. Direction. Arabic runs right to left, but the coordinates on the page are just numbers. If the parser rebuilds the line left to right, words come out reversed or letters come out in the wrong order.
3. Ligatures. Combinations like لا are frequently stored as a single glyph. On extraction that glyph may vanish, become a box, or split incorrectly.
4. Mixed direction. A line containing Arabic text, a Latin tool name, and a phone number contains three direction changes. This is where interleaving happens: مهندس مدني Civil Engineer 2019 شركة.

None of this is a limitation of Arabic. It is a consequence of how the file was produced.

The failure-mode table

Use this to diagnose whatever you are seeing in your own extraction test.

What you see	What caused it	The fix
Letters disconnected: ا ل ر ي ا ض	Glyph-level export, no text layer joins	Re-export from Word or Google Docs as PDF
Words reversed: ضايرلا	Direction rebuilt left to right	Re-export; avoid design tools for Arabic
Empty boxes or question marks	Font not embedded in the PDF	Use a standard system font, tick “embed fonts”
Nothing selectable at all	The page is an image (scan or flattened export)	Never scan a CV; export digitally
Arabic and English merged on one line	Two-column side-by-side layout	One column, one language per file
Dates missing from search	Arabic-Indic digits used	Convert all digits to 0-9
Contact details absent	They sit in a header, footer, or text box	Move them into the document body as plain text
Section headings ignored	Custom or decorative heading names	Use conventional headings: الخبرة العملية، التعليم، المهارات
Skills list empty	Skills placed inside a table or graphic	Plain comma-separated text lines

Fonts: the single biggest Arabic-specific cause

If you fix only one thing, fix the font.

Arabic display and calligraphic fonts are the most common source of unreadable extraction, because they are built for visual beauty, they carry heavy ligature tables, and they are frequently not embedded when the file leaves your machine. When a font is not embedded, the receiving system substitutes something else, and the mapping from glyph to character can break entirely.

Safe choices: Arial, Tahoma, Times New Roman, Traditional Arabic, Simplified Arabic, Noto Naskh Arabic. All standard, all widely available, all extract cleanly.

Avoid: anything described as a display, calligraphic, diwani, kufi, or handwriting font; anything you downloaded for a design project; and anything whose preview shows heavy decorative flourishes.

Two more font-level rules:

• No tatweel (kashida). The character used to stretch words visually — مـــهـــنـــدس — inserts extra characters into the extracted text and destroys keyword matching.
• No diacritics. Vowel marks are unnecessary in a professional document and add characters that may not match a plain-text search.

Digits: the quiet keyword killer

Arabic-Indic digits (٠١٢٣٤٥٦٧٨٩) and Western digits (0123456789) are different characters. They look like the same numbers to you and are entirely different to a search index.

If your dates read ٢٠٢١ – ٢٠٢٤ and a recruiter filters for candidates with experience since 2021, you do not match. Not because the system dislikes Arabic — because you and the filter are using different symbols.

So, across the whole file:

• Digits 0-9 only, in both languages.
• Gregorian dates, one consistent format: 03/2021 – 08/2024.
• If your certificate is Hijri, convert it and put the Hijri date in brackets if it genuinely matters.
• Phone numbers in +966 format, on their own line, not embedded in a sentence — mixed-direction phone numbers inside Arabic sentences are a common source of digit reordering.

Layout: the part that is not Arabic-specific but hits Arabic CVs harder

Arabic CV templates circulating online lean heavily on tables, boxes, sidebars, and mirrored two-column designs — far more than English templates do. So Arabic CVs fail more often, and people wrongly blame the script.

The rules are identical to any language:

• One column. No exceptions.
• No tables. Not for skills, not for languages, not for contact details.
• No text boxes or shapes. Text inside a shape frequently does not extract.
• No headers or footers holding anything you need read.
• Conventional section headings: الملخص المهني، الخبرة العملية، التعليم، المهارات، الشهادات والدورات.
• Real text, not an image. Never print, scan, and re-save. That produces a PDF containing zero characters.

The design-tool trap deserves its own warning: exports from graphic design tools regularly flatten text or store it in ways that survive printing but not parsing, and Arabic suffers worst because of shaping. Why Canva CVs fail ATS checks covers this in full.

Weak: a downloaded Arabic template with a coloured sidebar, icons beside each contact field, skills in a bordered table, and a decorative Diwani heading font.

Strong: a single-column Word document in Tahoma, plain conventional headings, contact details as three text lines at the top, exported to PDF with fonts embedded.

Bilingual CVs: what is safe and what is not

Three formats, one of which is genuinely dangerous.

Two separate files (best). Name-CV-AR.pdf and Name-CV-EN.pdf. Same facts, same dates, same numbers, each a clean single column. Send the one the employer wants; attach the second only when asked.

One file, stacked (acceptable). Arabic pages first, then English pages, in a single PDF. Use it only when the form accepts one attachment and you genuinely need both languages. The cost: the file doubles in length and the index treats both languages as one block, which dilutes keyword density.

Side-by-side columns (never). Arabic on the right, English on the left. The parser reads across the page and merges the two languages into lines nobody can use. This layout looks polished on paper and is the worst thing you can submit.

Terminology inside an Arabic CV

An Arabic CV does not mean everything is in Arabic. Professional vocabulary in the Saudi market is genuinely bilingual, and your file should work with that:

• Job titles: Arabic first, English in brackets on first use — محلل بيانات (Data Analyst).
• Tools and technologies: always Latin — Python, Power BI, SAP, AutoCAD. Nobody searches a CV database for “بايثون”.
• Certifications: official Latin names — PMP, CFA, CMA, SOCPA.
• Company names: written the way the company writes itself.
• File name: Latin characters. سيرة ذاتية.pdf breaks some upload forms and tells a recruiter nothing.

This bilingual layering is what lets one Arabic file match both Arabic and English search terms — which matters because recruiters in Saudi Arabia routinely search in English even when the role and the CV are Arabic. The same logic applies to profiles, covered in the keywords recruiters actually search.

The 60-second self-test

Do this before every submission. It is the closest thing to seeing your file through the system’s eyes.

1. Open your PDF in any reader.
2. Select all (Ctrl+A), copy (Ctrl+C).
3. Paste into a plain text editor — Notepad, TextEdit, or a blank text file. Not into Word, which will re-render the text and hide the problem.
4. Read what appears.

Then check, in order:

• Did anything paste? If not, your CV is an image and no system can read it.
• Are your name, phone, and email present? If not, they are in a header or a box.
• Is the Arabic readable as words, or is it disconnected letters and reversed strings?
• Do your dates appear as 2021, or as ٢٠٢١?
• Do job titles, employers, and dates appear in a sensible order, or are lines interleaved?
• Is every skill you listed present in the text?

Whatever fails here is what a recruiter’s search will never find. Fixing the export usually solves all of it at once: rebuild in Word or Google Docs, one column, standard Arabic font, then File → Save as PDF with font embedding on. Do not use “print to PDF” from a browser for an Arabic file.

What this does not fix

Parsing is a floor, not a ceiling. A perfectly readable Arabic CV still loses if the content is a list of duties instead of results, if the summary says nothing, or if the keywords do not match the posting. Clearing the machine only earns you the human, and the human reads for evidence: what changed, by how much, over what period, at what scale. See common CV mistakes for what gets files rejected after they parse correctly.

Rebuilding an Arabic file every time you find a new parsing problem is tedious, and the risk is that the Arabic and English versions drift apart until they disagree about your own career. TrueSira is built around that: one Master Profile holds your real experience once, and you paste a job description to derive a tailored, ATS-ready CV in whichever language the posting uses — clean single-column output, consistent digits, and every line tracing back to something you actually did, with you approving each one. Get started free.

FAQ

Do applicant tracking systems support Arabic?

Yes. The systems used by large Saudi employers handle Arabic text and right-to-left direction without difficulty, and Arabic CVs are processed every day. When an Arabic CV fails, the cause is almost always how the PDF was produced — a font that was never embedded, a design-tool export that stored glyph shapes, a scanned page, or a multi-column layout — not the script itself.

Why does my Arabic CV appear scrambled after upload?

Usually one of two causes. Either the file stores shaped glyph forms rather than clean logical text, so extraction produces disconnected or reversed letters, or the parser flattened a multi-column layout into a single reading order and merged unrelated content onto the same line. Re-exporting a single-column document from Word with a standard embedded Arabic font resolves most cases.

Which font should I use for an Arabic CV?

Use a standard system font — Arial, Tahoma, Times New Roman, Traditional Arabic, Simplified Arabic, or Noto Naskh Arabic — and make sure fonts are embedded when you export the PDF. Avoid decorative, calligraphic, Diwani, Kufi, and downloaded display fonts, which carry heavy ligature tables and are the most frequent Arabic-specific cause of unreadable text extraction.

Should Arabic dates use Western numerals?

Yes. Use digits 0-9 and Gregorian dates in one consistent format throughout the file. Arabic-Indic digits such as ٢٠٢٤ are a completely different set of characters, so a filter searching for 2024 will not match them even though they look like the same number to you. If a credential is Hijri, convert it and add the Hijri date in brackets only if it matters.

Is a bilingual CV safe for ATS?

A single-column file that switches language between sections parses fine, and stacking Arabic pages before English pages in one PDF is acceptable when only one attachment is allowed. What is never safe is a side-by-side two-column layout with Arabic on one side and English on the other, because the parser reads across the page and merges both languages into unusable lines.

How can I test whether my Arabic CV parses correctly?

Open the PDF, select all, copy, and paste into a plain text editor rather than into Word. What appears is very close to what the system extracts. Check that your name and contact details are present, that Arabic reads as connected words rather than reversed or disconnected letters, that dates show as 2024 rather than ٢٠٢٤, and that no lines have been interleaved.