Detect language API

detect_language takes a text and returns a list of the most likely languages.

It is built on top of the open-source Compact Language Detector 2 (CLD2) library.

Command name: detect_language Success status: 200 OK

On this page

• Common parameters

• Parameters

• Behavior notes

• Response schema

• Request examples

• Errors

Related: Overview, Supported languages, Why WProofreader auto detect language doesn't work for small piece of texts?

Common parameters

cmd selects the command to run.

These parameters work with any cmd.

See Overview for details.

Parameters

These are parameters specific to cmd=detect_language.

Behavior notes

• locale affects only LangName. It does not affect detection.

• If the language cannot be detected, you still get 200 OK.

  • The response contains a single item with LangShortCode=Unknown.

• For better accuracy on long texts, split content into sentences.

  • Run detect_language per sentence.

• At least 3 words in text is needed to detected the language. See the information here.

Response schema

The response contains language candidates.

Use format to switch between JSON and XML.

• LangShortCode (string) Detected language code. See Supported languages. Can also be Unknown.

• LangName (string) Human-readable language name. Localized using locale when provided.

• Proportion (number) Confidence / proportion of this language in the text. Higher means “more likely”.

Request examples

Example: single language detected

Request:

Response (200 OK):

Example: localized language name (locale=es-ES)

Request:

Response (200 OK):

Example: no language detected

Request:

Response (200 OK):

Errors

This command has no command-specific 4xx errors.

When language detection fails, the response is still 200 OK with LangShortCode=Unknown.

Common API errors (for example an invalid serviceid) use the shared error format from Overview.