SEO Tools for Arabic Keywords: How 8 Tools Handle Spelling Variants

Keyword tools disagree about Arabic spelling, and the disagreement changes the numbers. On 19 October 2026 we ran 40 Arabic keywords, 20 pairs that differ by one spelling habit, through eight SEO tools in the Saudi, UAE and Egyptian databases. Semrush showed a volume for both spellings in 13 of the 20 pairs. Google Ads figures, read through WordStream’s free tool, gave both spellings an identical volume and bid range in 8 of 19 spelling pairs.
None of this says which tool is right, because we did not check any figure against Search Console. It shows you whether SEO tools for Arabic keywords list a second spelling on its own or hand it the first spelling’s number. Where two or more tools gave a keyword a volume, 12 of 32 keywords landed more than twice apart. The six widest gaps all involved a pair that one tool merged and another split.
How we tested 40 Arabic keywords in eight tools
We built 20 keyword pairs: 7 Saudi, 6 Emirati and 7 Egyptian. In 19, both strings mean the same thing and differ by one habit: hamza, taa marbuta, alef maqsura, digit style, the article al- or word spacing. The 20th is a control: the Egyptian and standard words for “cars” in “used cars for sale”, which no tool should merge.
| Market | Keyword (English meaning) | What changes between the two spellings |
|---|---|---|
| Saudi Arabia | “car insurance” | Hamza on the alef: ta’min or tamin |
| Saudi Arabia | “car rental” | Hamza on the alef: ta’jir or tajir |
| Saudi Arabia | “furniture moving company in Riyadh” | Taa marbuta or final haa: sharika or sharikah |
| Saudi Arabia | “hospital near me” | Alef maqsura or yaa: mustashfa or mustashfi |
| Saudi Arabia | “furniture moving in Dammam” | Al- kept or dropped: al-dammam or dammam |
| Saudi Arabia | “24-hour dental clinic in Riyadh” | Western or Arabic-Indic digits for 24 |
| Saudi Arabia | “personal status (family law) lawyer” | Taa marbuta or final haa: shakhsiya or shakhsiyah |
| UAE | “iPhone 17 Pro Max” | Western or Arabic-Indic digits for 17 |
| UAE | “apartments for rent in Dubai” | Hamza below the alef: lil-‘ijar or lil-ijar |
| UAE | “Dubai visa” | Hamza on the alef: ta’shirat or tashirat |
| UAE | “cosmetic clinic in Dubai” | Taa marbuta or final haa: ‘iyada or ‘iyadah |
| UAE | “health insurance in the UAE” | Al- on two words kept or dropped: al-ta’min al-sihhi or ta’min sihhi |
| UAE | “jobs in Abu Dhabi” | One word or two: abuzabi or abu zabi |
| Egypt | “gold prices today in Egypt” | Hamza on the first alef, written or left out |
| Egypt | “apartments for sale in the Fifth Settlement (New Cairo)” | Fi (“in”) ending in yaa or in alef maqsura |
| Egypt | “applying for social housing apartments” | ‘Ala (“on”) ending in alef maqsura or in yaa |
| Egypt | “university admissions coordination 2026” | Western or Arabic-Indic digits for 2026 |
| Egypt | “restaurants in Alexandria” | Al- kept or dropped: al-iskandariya or iskandariya |
| Egypt | “shipping company” | Taa marbuta or final haa: sharika or sharikah |
| Egypt | “used cars for sale” (control) | Two words for “cars”: Egyptian ‘arabiyat or standard sayyarat |
We ran all 40 strings in Semrush, on a paid Semrush One Pro+ plan, and in WordStream’s free tool. We got 18, 10 and 7 strings through the free tiers of SE Ranking, Ahrefs and Keyword Tool before a cap or a verification page stopped us, and 3 through Backlinko. AnswerThePublic and Ubersuggest allowed one search each, and Moz asked for a login or a trial.
Google Keyword Planner is missing too. The Google Ads account available on the test day showed a billing notice when Keyword Planner opened, and we stopped there. WordStream labels its figures with the Google Ads logo, so it stands in as the Google-derived source, even though the tool is not Keyword Planner. With no Search Console anchor, we measure behaviour, not accuracy.
Do keyword tools keep two Arabic spellings apart?
Semrush and SE Ranking mostly list two Arabic spellings separately; Google Ads data merges them in more than half the pairs where it shows both. In our test, Semrush showed different volumes in 9 of the 12 spelling pairs where it had both, SE Ranking in all 5 of its covered pairs, and WordStream in only 7 of 16.
Take “car insurance” in Saudi Arabia. Semrush gave the hamza spelling, ta’min, 27,100 monthly searches and tamin 12,100, and SE Ranking split it 47,500 to 27,100. WordStream showed 135,000 for both, with the same $0.20 to $1.01 top-of-page bid range, which we read as one grouped entry.
Keyword Planner groups close variants by design. In a 2021 study of 72,635 keywords, Ahrefs found it folding misspellings such as “ahref” and “hrefs” into the figure for “ahrefs”. Ahrefs also says most SEO tools take their volumes from Keyword Planner. Tools that split a pair, Ahrefs among them, estimate each spelling from other data such as clickstream.
No tool merged the control pair. The Egyptian and standard words for “cars” got 6,600 and 18,100 searches in Semrush and in WordStream, and 1,800 and 18,100 in SE Ranking. WordStream did give both words the same bid range, $0.01 to $0.06, so matching bids alone do not prove a merge.
Google Ads data merges hamza pairs and splits digit pairs
WordStream’s Google Ads figures gave both spellings the same volume in all 5 hamza pairs we ran and in 2 of the 4 taa marbuta pairs. Digit pairs never merged. “iPhone 17 Pro Max” drew 22,200 searches with Western digits and 1,600 with Arabic-Indic ones. Semrush split most of the pairs it covered, whatever the habit.
Semrush split 3 of the 5 hamza pairs, and the other 2 rest on partial rows that copy the other spelling’s figures. Its hamza gaps run both ways: “car rental” drew 201,000 searches with the hamza and 5,400 without, while “gold prices today in Egypt” drew 12,100 with it and 246,000 without.
The other habits gave mixed results. Google Ads data merged the two spellings of ‘ala (“on”) but split fi (“in”). “Apartments for sale in the Fifth Settlement” drew 390 searches with the standard fi and 590 with the alef maqsura form common in Egypt. It merged the two-word al- pair and split the one-word pair, 110 against 590 for “restaurants in Alexandria”.
| Spelling change | Pairs | Semrush Keyword Overview (volumes) | Google Ads data via WordStream (volumes) | Semrush Keyword Magic Tool, Broad Match (idea list) |
|---|---|---|---|---|
| Hamza on or below the alef | 5 | 3 different, 2 identical (copied partial rows) | 5 identical (4 with identical bids) | Same list for both spellings in 5 of 5 |
| Taa marbuta or final haa | 4 | 1 different, 1 identical (copied partial row), 2 one spelling only | 2 identical, 1 different, 1 one spelling only | Same list in 4 of 4 |
| Alef maqsura or yaa | 3 | 1 different, 1 one spelling only, 1 neither | 2 different, 1 identical (‘ala) | Same list in 1 of 3 (fi) |
| Western or Arabic-Indic digits | 3 | 2 different, 1 one spelling only | 2 different, 1 one spelling only | Same list in 3 of 3 |
| Al- kept or dropped | 3 | 1 different, 2 one spelling only | 1 different, 1 identical (two-word al-), 1 one spelling only | Same list in 2 of 3 (one-word al-) |
| Word spacing | 1 | Different (590 and 50) | Different (720 and 70) | Different lists |
| Control: two words for “cars” | 1 | Different (6,600 and 18,100) | Different (6,600 and 18,100), identical bids | Different lists |
Semrush and Ahrefs return the same ideas for either spelling
Type either spelling into Semrush’s Keyword Magic Tool on Broad Match and you get the same keyword list in 15 of the 19 spelling pairs. Ahrefs’ free Keyword Generator returned identical top-20 lists, in the same order, for all 5 pairs it ran. Both tools fold the input together while still storing a separate volume for each spelling.
The folding has limits. Broad Match merged every hamza, taa marbuta and digit pair, the one-word al- pairs and the two forms of fi. It kept apart alef maqsura inside a word (“hospital”), ‘ala, the two-word al- pair and Abu Dhabi as one word or two. Phrase Match kept all 6 pairs we tried apart.
Broad Match also strips attached prefixes. Of the 91 Broad Match rows for “furniture moving company in Riyadh”, typed with the attached bi- (“in”), 53 contain that form and 36 use al-riyad without it. That widens your discovery but blurs the volume of the exact phrase you target. Site search engines fold Arabic letters in a similar way, as our guide to Arabic letter normalisation in on-site search covers.
Both tools fold the input together while still storing a separate volume for each spelling.
Keyword Tool’s autocomplete went the other way. Its totals changed with the spelling, 477 suggestions against 466 for “car insurance”, so it treats each spelling as its own query.
Why does one keyword show 12,100 in one tool and 673,000 in another?
The widest gaps appear where one tool splits a pair and another merges it. “Gold prices today in Egypt” with the hamza drew 12,100 monthly searches in Semrush, 165,000 in SE Ranking and 673,000 in WordStream’s Google Ads data, a 55.6x spread. WordStream gave the hamza-free spelling the same 673,000, while the other two tools put most searches there.
If you add the two spellings together, most of the gap closes for some pairs. SE Ranking’s two figures for that keyword sum to 615,000, close to WordStream’s 673,000, and Semrush’s 201,000 and 5,400 for “car rental” sum to 206,400, against WordStream’s 246,000. Other gaps stay open. Semrush’s two “car insurance” figures sum to 39,200, while WordStream shows 135,000.
Across our test, 32 strings had an exact volume in at least two tools. Half sat within 1.25x of each other, yet 12, or 37.5%, differed by more than 2x. Saudi strings spread widest, with a median high-to-low ratio of 2.00 and 4 of 9 beyond 2x. The UAE came in at 1.22 and 3 of 10, Egypt at 1.27 and 5 of 13.
Tools disagree on volume but agree on the split
When two tools both kept a pair apart, they gave the second spelling a similar share of the pair’s searches. Across the 7 such pairs in our test, the median gap between tools was 3.5 percentage points. “Apartments for sale in the Fifth Settlement” put 60.2% on the alef maqsura fi in all three tools that split it.
Agreement held at both ends of the scale. “iPhone 17 Pro Max” written with Arabic-Indic digits took 6.6% of its pair’s searches in Semrush and 6.7% in WordStream. “Restaurants in Alexandria” without al- took 80.8% and 84.3%, and “jobs in Abu Dhabi” with Abu Dhabi as two words took 7.8% and 8.9%.
Wider gaps appeared on “car insurance” without the hamza: 30.9% in Semrush, 36.3% in SE Ranking. “University admissions coordination 2026” in Arabic-Indic digits took 26.2% in Semrush and 20.2% in WordStream. One pair broke the pattern. In “gold prices today in Egypt”, the hamza-free spelling took 95.3% of the pair in Semrush and 73.2% in SE Ranking.
Tools that disagree on how many people search can still agree on how those searches divide between two spellings. Use the share to pick the spelling that leads your title or ad.
What Semrush shows when a spelling is missing
Semrush stored full metrics for 26 of the 40 strings, a partial row for 6 and nothing for 8, where a partial row shows volume, CPC and competition but no difficulty or intent. In 3 of the 6, those figures matched the other spelling’s row exactly, so those identical pairs tell you nothing about demand for the second spelling.
The clearest example is “apartments for rent in Dubai” with the hamza below the alef. Keyword Overview printed 1,600 searches and a $0.65 CPC, the same as the hamza-free spelling, yet that form appeared in none of its own Keyword Magic Tool lists. When we refreshed the 14 strings without full metrics, Semrush added difficulty, intent and SERP features, never a volume.
Six of our pairs showed only one spelling in Semrush. “Furniture moving in Dammam” drew 2,900 searches with al- and nothing without it. “Hospital near me” had no metrics in the Saudi database in either spelling, yet the standard spelling has a full row in the UAE database, at 1,000. A blank row tells you what the tool lacks; it says nothing about what searchers type.
A blank row tells you what the tool lacks; it says nothing about what searchers type.
The numbers also move. We pulled three of these pairs on 1 October for our study of how hamza and taa marbuta split Saudi search demand. Since then, “car insurance” with the hamza rose 22.1% to 27,100, while tamin stayed at 12,100. “Furniture moving company in Riyadh” went from an identical 12,100 and $4.20 CPC for both spellings to 12,100 and 14,800.
Before a keyword list sets a budget, both spellings need a number. Check both spellings of my top Arabic keywords
What free keyword tools show for Arabic
Every tool we ran accepted Arabic input and displayed it without right-to-left or broken-letter problems, including Arabic-Indic digits. The free tiers differ in what they show you and how soon they stop. WordStream ran all 40 strings with no cap or email request, while SE Ranking stopped at its limit of 5 analyses a day.
Ahrefs’ Keyword Generator shows volume buckets such as “>1000” for its first 20 ideas, and Keyword Tool blurs volume, CPC and competition. Both stopped us at Cloudflare checks after 8 to 10 searches. SE Ranking’s cap counts runs rather than keywords, and its unregistered bulk view shows 10 rows.
Three tools need care with the market. Backlinko’s free tool runs on Semrush data but has no country selector, and it returned Semrush’s US figures exactly. That meant 70 searches for “car insurance” with the hamza, against 27,100 in the Saudi database. Ubersuggest set the location to Egypt for a Saudi keyword and showed the CPC in pounds sterling. AnswerThePublic has no Arabic results language, though its one free search matched WordStream’s 135,000 for “car insurance”.
| Tool and tier | Arabic strings run | Volume shown as | Keyword ideas | Limit or block on 19 Oct 2026 |
|---|---|---|---|---|
| Semrush, paid (Semrush One Pro+) | 40 | Exact, with CPC | Median 268.5 per seed (Saudi 170, UAE 241, Egypt 1,178) | None; Saudi, UAE and Egyptian databases fully open |
| WordStream Free Keyword Tool (Google Ads data) | 40 | Exact, with top-of-page bid range | None beyond the typed keyword | No cap, no email; one keyword per search |
| SE Ranking free tools | 18 | Exact, displayed rounded (47.5K) | 99 similar keywords for one seed | 5 analyses a day; 10 rows in unregistered bulk results |
| Ahrefs free Keyword Generator | 10 | Buckets (>10,000, >1000, >100, <100) | First 20 shown, out of 54 to 2,158 | Cloudflare check after the 10th search |
| Keyword Tool, free | 7 | Hidden (blurred) | 104 to 668 suggestions (median 328) | Cloudflare check on the 9th search |
| Backlinko free keyword tool | 3 | Exact, Semrush US database only | Not counted | 3 checks a day; no country selector |
| AnswerThePublic | 1 | Exact | Autocomplete ideas and questions | 1 free search a day; no Arabic results language |
| Ubersuggest | 1 | Exact, for the location it set (Egypt) | Locked behind a free trial | 1 unregistered search; location set by IP |
| Moz Keyword Explorer | 0 | Not tested | Not tested | Login or free trial required |
| Google Keyword Planner | 0 | Not tested | Not tested | Billing notice in the available Google Ads account |
Can any tool claim accuracy for Arabic keywords?
We found no published study showing any tool is accurate for Arabic keywords. The studies we found test English keywords against Search Console, and none covers an Arab database. Semrush’s 2022 study used 10,000 US keywords, and its volume update reached 12 databases, none Arab. Ahrefs found Keyword Planner within 50% of Search Console for 45.22% of 72,635 keywords.
Those studies use Search Console as the benchmark because it counts impressions instead of estimating them, though only for queries where your own page appears.
Index size is a weak guide too. Ahrefs lists 2.7 billion keywords for the US. Its Arab indexes are far smaller: 445 million for Egypt, 433 million for the UAE, 389 million for Saudi Arabia, 327 million for Morocco and 308 million for Algeria. Semrush has 18 mobile databases, none Arab, and its SEO Writing Assistant supports 17 languages besides English, without Arabic. None of this tells you how a tool treats a hamza.
How to check an Arabic keyword before it reaches a budget
Run every money keyword in both spellings, in two kinds of tool, before it goes into your forecast. One tool should list spellings separately, such as Semrush’s Keyword Overview or SE Ranking. The other should show Google Ads figures. Then read the pair: identical, split, one-sided and missing results each mean something different.
- Write both spellings of each of your keywords, checking the six habits in this test: hamza, taa marbuta, alef maqsura, digits, al- and spacing.
- Run both in the right country database of a tool that stores spellings separately.
- Run both in Google Ads data, through Keyword Planner or WordStream’s free tool, and compare the bids with our Saudi and UAE CPC benchmarks.
- If Google shows the same volume and bids for both spellings, treat that figure as one pool and divide it by the splitting tool’s share.
- If a spelling has no data, mark it unknown rather than zero, and check the other country database and your Search Console queries.
- Where two tools differ by more than 2x, as 12 of our 32 strings did, anchor on Search Console impressions for queries where your page holds a top-three position.
- Lead titles, headings and Arabic URL slugs with the spelling that takes the larger share, and use the other where it reads naturally in the copy.
Our read
Choose an Arabic keyword tool for how it treats spelling and whether it opens your country database in full, never for the size of its keyword index. In our test, one keyword ranged from 12,100 to 673,000 monthly searches across three tools. The highest figure came from the tool that gave both spellings one number.
We would not plan Arabic search on one tool’s numbers. Google Ads data answers how big a pool of searches is, with spellings folded in, while a tool that splits spellings answers how that pool divides. In our test, the second answer held steadier between tools: a 3.5-point median gap on shares, against volume gaps above 2x on 12 of 32 strings.
Treat identical numbers for two spellings as one pool, and a missing spelling as unknown. Neither tells you that the second spelling has no searchers.
Treat identical numbers for two spellings as one pool, and a missing spelling as unknown.
Free tiers can flag a merged pair for you, but they cannot carry a market plan. Apart from WordStream, which returns no keyword ideas, none of the free tools ran more than 18 of our 40 strings before a cap or a check stopped it.
FAQ
Is Semrush accurate for Arabic keywords?
No published test answers that for Arabic. Semrush’s 2022 accuracy study used 10,000 US keywords, and its volume update reached 12 databases, none Arab. In our October 2026 test, Semrush showed a volume for 32 of 40 Arabic keywords and for both spellings in 13 of 20 pairs. Check it against your Search Console impressions.
Why do two Arabic spellings show the same search volume?
Google groups close variants into one entry, and tools that show Google Ads data repeat that entry for each spelling. In our test, WordStream’s Google Ads figures gave both spellings the same volume and bid range in 8 of 19 pairs, including 135,000 for “car insurance” with and without hamza. Treat such a pair as one pool.
Semrush shows no volume for my Arabic keyword. Why?
Semrush only stores metrics for keywords in its database, and some Arabic spellings are missing. In our test, 8 of 40 strings had no stored metrics, and refreshing them added difficulty and SERP features but never a volume. Check the other spelling of your keyword first: “furniture moving in Dammam” showed 2,900 searches with al- and nothing without it.
Which free keyword tool works with Arabic?
Every free tool we ran accepted Arabic, but each has a limit. WordStream gave Google Ads volumes for all 40 strings with no cap, one keyword per search and no ideas. Ahrefs’ Keyword Generator showed only volume buckets and stopped at a verification page after 10 searches. SE Ranking allowed 5 analyses a day, and Keyword Tool hid volumes.
Should I target the hamza spelling or the one without?
Decide per keyword, with a tool that lists both spellings. In Semrush, “car rental” in Saudi Arabia drew 201,000 searches with hamza and 5,400 without, while “gold prices today in Egypt” drew 246,000 without hamza and 12,100 with it. Lead your titles with the larger spelling and use the other where it reads naturally.
Which spelling decisions is your keyword tool making for you?
No published test says which tool is most accurate for Arabic. You can still learn what each tool does with a second spelling before it shows a number. In our test, the widest gaps sat exactly where one tool split a pair and another merged it. Once you know a tool’s habit, two figures 55 times apart stop looking contradictory.
If your site already ranks for Arabic queries, export 12 months of Search Console queries, find both spellings of your top terms and compare their impressions with each tool’s figures. If you are starting from zero, run the two-spelling check above on your 20 most valuable keywords before any budget or content plan is fixed.
Send us the keywords your forecast depends on. Run the two-spelling check on my Arabic keywords
Everything else we have run on Research & Data
Written by whoever ran the work, not a content team
51 articles







