RulesActive

Review repeated words

word-repetition

Flags content words repeated in the text above the threshold configured for the genre.

Usage of this metric

When enabled and matches are found, the report contains one item with this metric_id in result.recommendations. Absence means no match was found, not that the metric was not evaluated.

Location in the resultresult.recommendations[]metric_id = "word-repetition"

Calculation method

Counts how often each lemma occurs, excluding auxiliary and copular lemmas, and flags noun, adjective, adverb or verb lemmas whose frequency exceeds the genre threshold.

Result interpretation

Each `occurrences` item identifies an affected span. `start` is inclusive and `end` exclusive, both measured in Unicode code points over the submitted text. `replacement`, `alternatives` and `edits` contain values only when a safe change is available. Absence means no match was found, not necessarily that the text is correct.

Limitations

This is a review signal, not an automatic correction or a complete assessment of text quality.

Dependencies

language: en
implementation_version: 1.2

Valid example

{
  "category": "lexical",
  "examples": null,
  "explanation": null,
  "guidance": [
    "Review the detected fragment in context."
  ],
  "id": "rec_1",
  "implementation_version": "1.2",
  "metric_id": "word-repetition",
  "metric_title": "Review repeated words",
  "occurrences": [
    {
      "alternatives": [],
      "edits": [],
      "end": 25,
      "id": "occ_1",
      "paragraph_index": 0,
      "replacement": null,
      "sentence_index": 0,
      "start": 18,
      "text": "example"
    }
  ],
  "summary": "Example detection returned by this metric."
}

GET /v1/metrics/word-repetition

Complete request, response and error contracts, generated automatically from the code.

Open the JSON record