Keyword Grouper

What the tool does

Keyword research often ends with dozens of spellings of the same phrase: İSTANBUL otel, istanbul otel, istanbul otelleri. The tool splits such a list into groups and shows each group’s key, its members, the total volume if you supplied one, and why every member ended up in that group.

There are two views. The phrase view only merges variants that reduce to the same word sequence. The shared-word view builds topic clusters around the most common word that appears in at least two keywords. Keywords left on their own and undecided suggestions are listed separately.

How to use

1. Paste one keyword per line. To add volumes, paste tab-separated rows from a spreadsheet or switch to CSV with keyword,volume.
2. Choose how to group: phrase variants or shared-word clusters.
3. Optionally enable root/suffix suggestions and write your own alias rules such as otelleri => otel.
4. Press “Group”. Long lists are processed in the background and can be stopped with “Cancel”.
5. Rename a group or change a row’s group field, press “Apply edits”, then download the CSV.

Turkish normalization

Exact matching converts each line to NFC, lower-cases it with Turkish rules and collapses whitespace. İSTANBUL becomes istanbul and IŞIK becomes ışık; dotted and dotless i are never mixed up the way English rules would. Repeated lines are kept as one member, the reason shows how often they appeared, and the highest volume is used.

The alias dictionary works on single words: variant => root or v1, v2 => root. Word order, negation and purchase-intent words are never changed, so ucuz değil (“not cheap”) never merges with ucuz in any mode.

Root and suffix suggestions

Suggestions only run when the box is ticked, and they are deliberately cautious. The rule: remove one suffix layer, try the longest suffix first, and require the remaining root to appear in the list on its own. The suffix table covers the plural (-lar/-ler), -ları/-leri, locative, ablative, genitive, possessive and dative endings, checked against vowel harmony and d/t assimilation.

• kitap and kitaplar merge because kitap is in the list and -lar fits.
• kitabı merges only because the built-in consonant-alternation table states kitab → kitap; for other words you add an alias.
• yaz and yazar stay apart; there is no -ar suffix in the table.
• Single-vowel suffixes (as in kalemi) and words that yield more than one valid root are never merged; they go to the undecided list for you.

Example and interpretation

The 13-line example produces 8 groups, 4 of them with more than one member. kalemi lands in the undecided list because -i can be either an accusative or a possessive ending. In the shared-word view the same list forms a kitap cluster (6 members) and a kırtasiye cluster (4 members), while ucuz kitap değil stays in kitap + değil. Group volume is a plain sum; search engines treat close variants together, so real demand can be lower.

Limits of meaning and search intent

The tool looks at spelling, not meaning. It does not perform full Turkish morphological analysis, synonym detection, search-intent classification, SERP similarity or AI embeddings, and it does not connect to Google data. ankara kebap and ankara kebapçı may share an intent, but the tool cannot know that; your alias rules and edits make that call. Before turning clusters into a content plan, checking the search results for the important ones by hand is a good habit.

Limits and privacy

• Up to 5,000 lines and 200 characters per line are processed; longer input shows an error instead of being cut.
• The alias dictionary accepts up to 500 rules; conflicting rules are reported with their line number.
• CSV export escapes cells starting with =, +, - or @ so spreadsheets do not run them as formulas.
• The list is processed in your browser, never sent to a server and never written to storage. It only becomes part of a link if you choose to include it when sharing.