Discuss with expertTry it for free
© 2026 Direct Cursus Technology L.L.C.

OCR

Written by
Updated at August 18, 2026
View in Markdown

Note

Currently, there are two UIs that support Workflows: Yandex Cloud and Yandex AI Studio. In this case, the following applies:

  • Workflows created in the Yandex Cloud UI are automatically available in the AI Studio UI.
  • Workflows created in the AI Studio UI are not available in the Yandex Cloud UI.

Starting September 3, 2026, Workflows will no longer be supported in the Yandex Cloud UI. To create and manage workflows, use the AI Studio UI.

Image-to-text recognition with Yandex Vision OCR.

The base64, url, objectStorage, and disk fields are mutually exclusive: you can use only one of them.

Field name Type Required Default value Templating supported Description
recognizeText OCR.RecognizeText Yes None None Text recognition settings.
base64 string None None Yes Base64-encoded image.
url string None None Yes Image via a public URL.
objectStorage OCR.ObjectStorage None None None Image from an Object Storage bucket.
disk OCR.Disk None None None Image on Yandex Disk.

OCR.RecognizeText

Field name

Type

This is a required field

Default value

Templating supported

Description

mimeType

string

Yes

None

None

Image file format. Available formats: pdf, jpeg, png.

languageCodes

string[]

Yes

None

None

Text recognition languages.

Available languages
  • ru: Russian
  • en: English
  • ar: Arabic
  • az: Azerbaijani
  • ba: Bashkir
  • be: Belarusian
  • bg: Bulgarian
  • bs: Bosnian
  • cs: Czech
  • da: Danish
  • de: German
  • el: Greek
  • es: Spanish
  • et: Estonian
  • fi: Finnish
  • fr: French
  • he: Hebrew
  • hu: Hungarian
  • hy: Armenian
  • id: Indonesian
  • it: Italian
  • ja: Japanese
  • ka: Georgian
  • kk: Kazakh
  • ko: Korean
  • ky: Kirghiz
  • lt: Lithuanian
  • lv: Latvian
  • mt: Maltese
  • nl: Dutch
  • no: Norwegian
  • pl: Polish
  • pt: Portuguese
  • ro: Romanian
  • sk: Slovak
  • sl: Slovenian
  • sv: Swedish
  • tg: Tajik
  • th: Thai
  • tr: Turkish
  • tt: Tatar
  • uk: Ukrainian
  • uz: Uzbek
  • vi: Vietnamese
  • zh: Chinese

model

string

Yes

None

None

Text recognition model.

Available models
  • page: Text recognition on page.
  • page-column-sort: Column-sorted text recognition.
  • passport: Russian Federation passport.
  • driver-license-front: Driver's license, front side.
  • driver-license-back: Driver's license, back side.
  • license-plates: Car license plates.
  • vehicle-registration-front: Vehicle registration certificate (VRC), front side.
  • vehicle-registration-back: Vehicle registration certificate (VRC), back side.
  • handwritten: Handwritten text.
  • table: Table.
  • markdown: Markdown markup.
  • math-markdown: Markdown markup with formulas.

OCR.ObjectStorage

Field name Type Required Default value Templating supported Description
bucket string Yes None None Name of the Object Storage bucket the image file is in.
object string Yes None Yes Object name, e.g., prefix/subprefix/data.png.

OCR.Disk

Field name Type Required Default value Templating supported Description
oauthToken string Yes None None OAuth token for authentication when accessing Yandex Disk.
path string Yes None Yes Image file path.
sharedDiskId string None None Yes Shared disk label.
Previous
DatabaseQuery
Next
STT