# Overview

This reference describes RESTful web methods for interacting with Tisane API using standard HTTP and JSON-based communication. 

Version: 4.8.0.0

## Servers

Tisane public cloud
```
https://api.tisane.ai
```

## Security

### Tisane-API-Key

API subscription key (primary or secondary) from [Tisane Developer Portal](https://dev.tisane.ai/profile)

Type: apiKey
In: header
Name: Ocp-Apim-Subscription-Key

## Download OpenAPI description

[Overview](https://docs.tisane.ai/_bundle/apis/tisane-api-short.yaml)

## NLU / NLP Methods

These are the methods that actually perform analysis and transformation of input text.

### Analyze text

 - [POST /parse](https://docs.tisane.ai/apis/tisane-api-short/nlu-nlp-methods/parse.md): The method analyzes the input and detects problematic content, sentiment snippets, entities, topics, phrase structure, parts of speech, stopwords, and more.

### List available languages

 - [GET /languages](https://docs.tisane.ai/apis/tisane-api-short/nlu-nlp-methods/languages.md): Retrieve a list of supported languages. 

The response includes the following attributes:
* id - Language ID.
* name - Native name.
* englishName - English name.
* nativeEncoding - Encoding of the language.
* preferredFont - Recommended font for display.
* latin - Whether the language uses Latin script.
* rightToLeft - Whether the language uses right-to-left script (e.g., Arabic, Hebrew, Persian).

### Text clean-up

 - [POST /helper/extract_text](https://docs.tisane.ai/apis/tisane-api-short/nlu-nlp-methods/extract_text.md): A service method to remove JavaScript, CSS tags, JSON, and other markup, returning pure decoded text.
The request body includes the markup content from which text will be extracted.

Note: This method does not process binary content.

### Named entity comparison

 - [POST /compare/entities](https://docs.tisane.ai/apis/tisane-api-short/nlu-nlp-methods/compareentities.md): Compares two compound named entities and identifies differences.
The request body should include:
* language1 (string) - IETF tag for the first entity's language.
* entity1 (string) - The first entity.
* language2 (string) - IETF tag for the second entity's language.
* entity2 (string) - The second entity.
* type (string) - The entity type (currently only _person_ is supported).

More information: Named Entities in the Tisane API Response Guide
The response contains:
* result (string) - Comparison result:
  * no_single_entity - One or both entities are invalid.
  * same - Entities are identical (name order might differ).
  * different - Entities differ, with details in differences.

* differences (array of strings) - Lists detected differences:
  * given_name
  * surname
  * title (e.g., Mr., Mrs.)
  * social_role (e.g., academic degrees)
  * suffix
  * variation (e.g., spelling or style differences)

### Semantic similarity

 - [POST /similarity](https://docs.tisane.ai/apis/tisane-api-short/nlu-nlp-methods/similarity.md): Calculate the semantic similarity between two text fragments, either in the same language or in different languages.

The request body should contain:
* content1 (string) - The first text fragment.
* language1 (string) - The IETF language code for content1.
* content2 (string) - The second text fragment to compare.
* language2 (string) - The IETF language code for content2.
* settings (object) - Additional settings as per specifications. More information: Tisane API Configuration And Customization Guide.

The response is a number between 0 and 1, representing the similarity of the submitted text fragments.

### Detect language

 - [POST /detectLanguage](https://docs.tisane.ai/apis/tisane-api-short/nlu-nlp-methods/detectlanguage.md): Detects the languages used in the provided text fragment and returns the breakdown by offsets.

The request body should include:
* content (string) - the text fragment to analyze.
* languages (string, optional) - a vertical bar-delimited list of language codes to use as cues.
* delimiter (string, optional) - a regular expression for segmenting the fragment; by default, the fragment is not segmented.
The response is a JSON structure containing languages, which provides a breakdown of language codes.

### Translate text

 - [POST /transform](https://docs.tisane.ai/apis/tisane-api-short/nlu-nlp-methods/transform.md): This method translates the input text.
The request body should include the following elements:
 from (string) - A standard IETF tag for the source language. Use  or a vertical bar-delimited set of language codes to invoke autodetect.
* to (string) - A standard IETF tag for the target language. If the source and target languages are the same, paraphrasing will be applied.
* content (string) - The text content to translate.
* settings (object) - Additional translation settings.

More information: Tisane API Configuration And Customization Guide

The response is the transformed text.

## Language Model Direct Access

Methods to retrieve and inspect entries from the language models.

### List inflected forms

 - [GET /lm/inflections](https://docs.tisane.ai/apis/tisane-api-short/language-model-direct-access/inflections.md): Retrieves inflected forms of a specified lexeme within a given language family.