Keyword Extractor and Frequency Checker

1. Normalize letter case and curly apostrophes. Strip English possessive ’s. Do not stem words: “request” and “requests” remain separate.
2. Treat punctuation, digits, symbols and line breaks as boundaries. A hyphen separates phrases; words on opposite sides are not joined.
3. Count contiguous one-, two- and three-word spans that contain no configured English stopwords. Stopwords stay in the total word count and cannot connect two phrases.
4. Count overlapping occurrences. Counts are phrase occurrences, not unique sentences, documents or users.

Example: “Product feedback. Product feedback helps teams.” gives “product feedback” a count of 2. It never creates “feedback product” across the full stop. The complete teaching sample can be loaded in the tool.

Repeated phrases can expose the vocabulary of a permitted discussion or competitor description. Select candidates that describe an input, obstacle or output. Exclude navigation boilerplate and ambiguous terms. “Feedback” may be frequent but “weekly feedback digest” gives a more specific hypothesis.

Copy or export your candidates. A Research link only prefills the existing research entry; it never automatically creates a paid task. Use the SaaS workflow to group candidates and the long-tail selection guide to refine them.

Maximum input: 50,000 characters. The table displays at most 300 matching phrases; exports include all matches, or your selected matching shortlist. Text and results are cleared on reload and are not persisted. Use Clear to remove them immediately.

This is a plain-text English phrase counter, not semantic analysis, a search-volume database, keyword-density advice or an Amazon-specific tool. Paste rendered text, not HTML source. Counts do not predict search demand, rankings or a Google spam decision.