How Phrase Permutations Are Built
A permutation generator takes lists of words and joins them into phrases using fixed patterns. If you have three seeds and four words that sit before them, the modifier-before-keyword pattern produces twelve phrases. Nothing about those phrases is measured. The tool does not know whether anyone searches for them, how competitive they are, or what they mean to a reader.
Treat the output as a drafting aid. Good use is planning page topics, section headings or ad copy variations that you will review, edit and check against real research. Poor use is publishing every generated line, most of which will read awkwardly once they are in a sentence.
All of the work happens in your browser. The lists you type are read into memory, combined and shown on the page. The copy and CSV actions write local output, and nothing is stored between visits.
Choose modifiers from your own vocabulary: product features, questions customers actually ask, and the terms your team uses. Words copied from a competitor's page reflect their wording rather than yours, and they may carry meanings you do not intend.
Walkthrough of Every Control
- Seed keywords. Your core terms, one per line. Each seed is joined with every word that the chosen patterns use.
- Modifiers before and after. Two lists of words you write yourself. Before-words sit in front of the seed and after-words sit behind it. Leave a list empty and the patterns that need it produce nothing.
- Audience terms. Used only by the pattern that writes the seed, then the word for, then an audience term. Leave it empty if you do not need that pattern.
- Patterns. Four checkboxes. The modifier-plus-keyword pattern builds before-word and seed. The keyword-plus-modifier pattern builds seed and after-word. The for-audience pattern builds seed, for and audience. The modifier-keyword-modifier pattern uses both word lists around the seed. It grows quickly, so it starts unchecked.
- Remove duplicates. On by default. Two phrases that differ only in capital letters count as the same, and the first one stays.
- Lowercase output. Converts every phrase to lowercase. It is applied before duplicates are removed, so its effect on the count is limited to the output text.
- Maximum phrases. Caps the kept output between 1 and 5,000. The summary reports how many phrases the cap dropped.
- Generate, copy and CSV. Generate rebuilds the output from the current fields. Copy list places one phrase per line on the clipboard. Download CSV writes a file with two columns, phrase and pattern.
Each list accepts up to 300 lines. The combined number of candidate phrases is limited to 200,000, and above that the tool asks you to reduce the lists instead of slowing the page down.
For a first pass, use the first two patterns with a short seed list. Add the for-audience pattern only when the audiences are real groups you serve. Save the modifier-keyword-modifier pattern for last, after the shorter lists have been reviewed, because it multiplies quickly.
Worked Example: The Sample Lists
Load the sample and keep the first three patterns checked. The sample seeds are running shoes, trail shoes and running shoes, with the last one repeated on purpose. The before-words are best, Best, affordable and lightweight, with an empty row between the first and second. The after-words are reviews, buying guide and near me. The audience terms are beginners, marathon runners and wide feet. The results read:
- Modifier + keyword: 12 generated, which is three seed rows times four before-words.
- Keyword + modifier: 9 generated, which is three seeds times three after-words.
- Keyword + for + audience: 9 generated, which is three seeds times three audience terms.
- Modifier + keyword + modifier: not selected in this run.
The summary line reads Generated 30 phrases. Removed 12 duplicates. Kept 18. Skipped 1 empty line.
The kept list, in order, is:
best running shoes affordable running shoes lightweight running shoes best trail shoes affordable trail shoes lightweight trail shoes running shoes reviews running shoes buying guide running shoes near me trail shoes reviews trail shoes buying guide trail shoes near me running shoes for beginners running shoes for marathon runners running shoes for wide feet trail shoes for beginners trail shoes for marathon runners trail shoes for wide feet
The repeated seed and the capital Best both produce phrases that already exist, and the duplicate rule removes them. Turn off Remove duplicates and the same run keeps every candidate, which is useful only when you want to audit the raw count.
A useful next step is to download the CSV and filter by the pattern column. Each pattern can then get its own review rule, such as checking every for-audience phrase against the audience you actually serve.
Tips and Common Mistakes
- Start with a short list of strong seeds. Long seed lists multiply quickly and mostly produce weak phrases.
- Write your own modifier lists. Copying word lists from other sites makes the output reflect their business rather than yours.
- Read every phrase. Mechanical joins can produce awkward text, such as a seed followed by a qualifier that does not fit.
- Use audience terms only for audiences you actually serve, so the for-audience phrases describe real readers.
- Keep the duplicate rule on for planning. Duplicates double the review work without adding anything new.
- Do not treat the result as demand data. The tool cannot say whether a phrase is searched, and that needs research data from a dedicated source.
- Export the CSV and sort it in a spreadsheet by pattern. Review the group that matters most first.
- Review in batches of 50 to 100 phrases. Smaller batches are easier to judge consistently than one long list.
- Keep a short record of which phrases you chose and why, so the next round starts from decisions rather than from scratch.
Frequently Asked Questions
No. It builds text combinations from your lists and nothing more. Demand, competition and trends need data from a research source that measures them.
It multiplies the seeds by both word lists, so a short list can still produce a long output. Turn it on when you have trimmed the lists and want the extra combinations.
It caps how many phrases appear in the kept list and the CSV. The summary reports how many were dropped, so you know the full count before the cap.
A cell that starts with an equals sign, plus, minus, at sign or tab could be read as a formula. The apostrophe makes the spreadsheet show the text as written, and it is the only change made to the cell.