Text Lemmatizer Skill

Overview

Lemmatizer parses `text` with the compromise natural-language library and asks it to compute a root for every term. compromise tags each word with its part of speech and uses that to find the dictionary form, which is why irregular forms such as "geese" and "went" resolve to "goose" and "go". Before parsing, the first tab, carriage return and line feed in the text are replaced with spaces and the text is trimmed.

For each term the normalized word is checked against a short stopword list (the, a, an, in, on, at, to, from, for, with, by, of), and those words are skipped. The root has any trailing possessive "'s" removed. Because compromise maps forms of "be", "have" and "do" inconsistently, the service applies its own table on top: is, am, are, was, were, been and being become `be`; has, had and having become `have`; does, did and done become `do`. Each resulting lemma is counted, and different surface forms of the same word add to the same count.

`lemmas` is an object keyed by lemma with an integer occurrence count as the value, and `found` is the number of keys. The object is capped at the first 50 distinct lemmas in the order they first appear; later ones are dropped and `found` reports the capped total. Words that are already in base form are included, and words compromise cannot resolve fall back to their normalized spelling, so some past-tense forms can appear unchanged. If no lemma is produced, the request returns a 404. On the Free plan `lemmas` is null and only `found` is returned.

Live Test Text Lemmatizer Skill Skill →

The tool

Once your client is connected to the VerveKit server, this appears in its tool list as TextLemmatizerSkill. It is read-only and open-world — it fetches and never mutates anything on your side — so most clients call it without asking you to confirm.

{
  "name": "TextLemmatizerSkill",
  "arguments": {
    "text": "The dogs were running and jumping over the fences"
  }
}

You do not name the tool yourself; the model picks it. Asking about The dogs were running and jumping over the fences in the terms this skill covers is enough for it to reach for TextLemmatizerSkill on its own — naming it explicitly also works, and is the way to force the call.

Connecting

One server URL covers every skill in the catalog, including this one. Authorization is OAuth: the client opens a browser once, and there is no key to paste into a config file.

{
  "mcpServers": {
    "vervekit": {
      "url": "https://api.vervekit.com/v1/mcp"
    }
  }
}

https://api.vervekit.com/v1/mcp

Per-client setup — Claude, Cursor, VS Code, ChatGPT — is on the MCP setup page.

Arguments

These are the properties on the tool's inputSchema, so a well-behaved client validates them before the call is made. Premium arguments are accepted on every plan but only take effect on plans that include them.

What the model gets back

The result carries a structuredContent object matching the tool's declared outputSchema, so a client reads fields without parsing prose. status is "ok" and error is null on success; a null field means the value was not available for that input, not that the call failed.

1
2
3
4
5
6
7
8
9
10
11
12

{
  "status": "ok",
  "error": null,
  "data": {
    "found": 3,
    "lemmas": {
      "cat": 1,
      "ran": 1,
      "door": 2
    }
  }
}

Response fields

Paths are relative to data. Premium fields are absent rather than zeroed on plans that do not include them, so check for presence instead of comparing to 0.

Failure modes

Errors come back as tool errors carrying a sentence the model can act on, not a bare status code. Error handling covers the full list.

A call costs 2 credits each time the tool actually runs; a model that reasons about the tool without calling it costs nothing.

Use cases

Search Query Normalization

Search engines convert user queries into root words to match plural or conjugated variations against indexed product titles.

Customer Feedback Clustering

Group incoming support tickets by topic after reducing verbs and nouns to base forms for cleaner text classification.

Vocabulary Learning Systems

Language learning platforms detect inflected verb forms across reading exercises and map them back to headwords in student flashcards.

SEO Content Auditing

Before scanning articles for keyword density, content marketing platforms strip grammatical suffixes to measure true term frequencies across drafts.

Other ways to use Text Lemmatizer Skill

Set up Text Lemmatizer Skill on VerveKit, or reach the same source a different way. Your VerveKit account and credits work on all of them — one key, one balance.

Related

More in Text Processing: