LEMMA-IS(1)

lemma-is - Icelandic word lemmatization service

curl -X POST https://lemma.solberg.is/api/lemmatize \
  -H "Content-Type: application/json" \
  -d '{"word":"börnin"}'

A free public API for Icelandic lemmatization. Maps inflected Icelandic word forms to their base forms (lemmas). Useful for search indexing, text analysis, and NLP tasks.

Powered by lemma-is npm library.

Method	Path	Description
GET	/	This documentation
GET	/health	Health check
POST	/api/lemmatize	Single word → lemmas
POST	/api/lemmatize/morph	Word → lemmas with morphology
POST	/api/text	Text → unique lemmas
POST	/api/batch	Multiple words (up to 1000)

Single word:

curl -X POST https://lemma.solberg.is/api/lemmatize \
  -H "Content-Type: application/json" \
  -d '{"word":"börnin"}'

# Response: {"word":"börnin","lemmas":["barn"]}

With morphological info:

curl -X POST https://lemma.solberg.is/api/lemmatize/morph \
  -H "Content-Type: application/json" \
  -d '{"word":"hestinum"}'

# Response: {"word":"hestinum","lemmas":[{"lemma":"hestur","category":"no","gender":"kk","case":"þgf","number":"et"}]}

Text indexing:

curl -X POST https://lemma.solberg.is/api/text \
  -H "Content-Type: application/json" \
  -d '{"text":"Börnin fóru í skólann"}'

# Response: {"original":"Börnin fóru í skólann","lemmas":["barn","fara","fóra","skóli","í"]}

Batch processing:

curl -X POST https://lemma.solberg.is/api/batch \
  -H "Content-Type: application/json" \
  -d '{"words":["hestur","hestinn","hestinum"]}'

# Response: {"results":[{"word":"hestur","lemmas":["hestur"]},{"word":"hestinn","lemmas":["hestur"]},{"word":"hestinum","lemmas":["hestur"]}]}

Code	Meaning	English
Case (fall)
nf	nefnifall	nominative
þf	þolfall	accusative
þgf	þágufall	dative
ef	eignarfall	genitive
Gender (kyn)
kk	karlkyn	masculine
kvk	kvenkyn	feminine
hk	hvorugkyn	neuter
Number (tala)
et	eintala	singular
ft	fleirtala	plural

Limit	Value
Word length	100 characters
Text length	50,000 characters
Batch size	1,000 words
Rate limit	100 requests/minute per IP

API Key: To bypass rate limiting, include X-API-Key: <key> header.

Disambiguation: The /api/text endpoint uses context-aware disambiguation. For example, “u00e1” is resolved to the verb “eiga” (to own) when preceded by a subject noun, or kept as a preposition when followed by a dative noun. This uses grammar rules, bigram frequencies, and suffix-based case inference.

Stopwords: The API does not filter stopwords — all lemmas are returned. The underlying lemma-is npm library supports contextual stopword removal (e.g., “u00e1” as preposition = stopword, “u00e1” as noun “river” = kept). Use the library directly with removeStopwords: true and useContextualStopwords: true for search indexing.

github.com/jokull/lemma-is - npm library

BÍN - Beygingarlýsing íslensks nútímamáls

Jökull Sólberg <jokull@solberg.is>