REST API · v0.2 · Cloud Run

Deterministic Kazakh stemmer for search normalization

Reduces word forms to their root for full-text search, indexing, and NLP. Every result is validated against a dictionary of 140,000+ lemmas — the stemmer never returns a nonexistent root.

Try in playground Quick start

0.0000

Over-stemming Index

140k+

Lemmas in dictionary

~10k/с

Words per second

Playground

Enter Kazakh text — the API returns roots in real time. No signup required.

input

Examples:

POST /stem/batch

// the result will appear here

● dictionary — root is in the dictionary● exact — whole word is in the dictionary● guess — rule-based, no dictionary match● unknown — not recognized

Quick start

One POST request. No keys, no SDK — fits any stack.

curl -X POST https://kazakh-stemmer-590833642796.europe-west1.run.app/stem/batch \
  -H "Content-Type: application/json" \
  -d '{"words": ["балаларымызда", "кітабым"]}'

Interactive Swagger documentation: https://kazakh-stemmer-590833642796.europe-west1.run.app/docs