LINDAT Korektor API

# Provenance: DERIVED by API Evangelist from the operator's own published REST API
# reference, then verified against live responses from the production host.
# UFAL publishes human documentation, not a machine-readable contract, for this service.
# method: derived
# x-source-url: https://lindat.mff.cuni.cz/services/korektor/api-reference.php
# generated: 2026-09-01
# x-operator: institution (Institute of Formal and Applied Linguistics, Faculty of
#   Mathematics and Physics, Charles University — host lindat.mff.cuni.cz is the
#   institution's own registrable domain)
openapi: 3.2.0
info:
  title: LINDAT Korektor API
  version: '1'
  summary: Czech spellchecking, diacritics generation and correction-suggestion service operated by ÚFAL at Charles University.
  description: >-
    Public REST API for Korektor, a statistical spellchecker and diacritics generator operated by
    the Institute of Formal and Applied Linguistics (ÚFAL), Faculty of Mathematics and Physics,
    Charles University, and delivered through the LINDAT/CLARIAH-CZ research infrastructure.
    Auto-corrects text or returns ranked spelling suggestions per token. No API key is required.
    Derived by API Evangelist from the published REST API reference at
    https://lindat.mff.cuni.cz/services/korektor/api-reference.php and verified against live
    responses from https://lindat.mff.cuni.cz/services/korektor/api on 2026-09-01.
  contact:
    name: LINDAT/CLARIAH-CZ Helpdesk
    email: lindat-help@ufal.mff.cuni.cz
    url: https://lindat.mff.cuni.cz/user_feedback
  license:
    name: CC BY-NC-SA
    url: https://creativecommons.org/licenses/by-nc-sa/4.0/
  termsOfService: https://lindat.mff.cuni.cz/en/terms-of-use
servers:
  - url: https://lindat.mff.cuni.cz/services/korektor/api
    description: Production service operated by ÚFAL, Charles University
tags:
  - name: korektor
    description: Spellchecking and diacritics operations
paths:
  /models:
    get:
      operationId: listKorektorModels
      summary: List available Korektor models
      description: Returns the array of available model names and the default model, guaranteed to be the latest Czech spellchecking model.
      tags: [korektor]
      responses:
        '200':
          description: Model inventory
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ModelList'
  /correct:
    get:
      operationId: correctText
      summary: Auto-correct the supplied text
      tags: [korektor]
      parameters:
        - $ref: '#/components/parameters/Data'
        - $ref: '#/components/parameters/Model'
        - $ref: '#/components/parameters/Input'
      responses:
        '200':
          description: Corrected text
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/CorrectResponse'
  /suggestions:
    get:
      operationId: suggestCorrections
      summary: Generate ranked spelling suggestions for the supplied text
      tags: [korektor]
      parameters:
        - $ref: '#/components/parameters/Data'
        - $ref: '#/components/parameters/Model'
        - $ref: '#/components/parameters/Input'
        - name: suggestions
          in: query
          required: false
          description: Maximum number of suggestions returned for a single token.
          schema:
            type: integer
            minimum: 1
            default: 5
      responses:
        '200':
          description: Suggestion result
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/SuggestionsResponse'
components:
  parameters:
    Data:
      name: data
      in: query
      required: true
      description: Input text in UTF-8.
      schema:
        type: string
    Model:
      name: model
      in: query
      required: false
      description: Model to use, for example `czech-spellchecker`, `czech-diacritics_generator` or `strip_diacritics`.
      schema:
        type: string
    Input:
      name: input
      in: query
      required: false
      description: Input format to use.
      schema:
        type: string
        enum: [untokenized, untokenized_lines, segmented, vertical, horizontal]
        default: untokenized
  schemas:
    ModelList:
      type: object
      properties:
        models:
          type: array
          items:
            type: string
        default_model:
          type: string
      required: [models, default_model]
    CorrectResponse:
      type: object
      properties:
        model:
          type: string
        acknowledgements:
          type: array
          items:
            type: string
            format: uri
        result:
          type: string
          description: The corrected text.
      required: [model, result]
    SuggestionsResponse:
      type: object
      properties:
        model:
          type: string
        acknowledgements:
          type: array
          items:
            type: string
            format: uri
        result:
          type: array
          description: Array of suggestion groups. Each group is an array of strings whose first element is the original span and whose remaining elements are suggestions ordered most to least probable. Concatenating the first elements reproduces the input.
          items:
            type: array
            items:
              type: string
      required: [model, result]