STT

Written by

Updated at August 4, 2026

View in Markdown

STT.RecognizeText
STT.ObjectStorage
STT.Disk

Speech recognition in audio files using Yandex SpeechKit STT.

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`	STT.RecognizeText	Yes	No	No	Speech recognition settings.
`base64`	`string`	No	No	Yes	Base64-encoded audio file.
`url`	`string`	No	No	Yes	Audio file via a public URL.
`objectStorage`	STT.ObjectStorage	No	No	No	Audio file from an Object Storage bucket.
`disk`	STT.Disk	No	No	No	Audio file on Yandex Disk.

STT.RecognizeText

Field name	Type	This is a required field	Default value	Templating supported	Description
`mimeType`	`string`	Yes	No	No	Audio file format. Available formats: `mp3`, `wav`, `ogg`.
`languageCodes`	`string[]`	Yes	No	No	Speech recognition languages. Available languages `auto`: Automatic language detection `ru-RU`: Russian `en-US`: English `de-DE`: German `es-ES`: Spanish `fi-FI`: Finnish `fr-FR`: French `he-IL`: Hebrew `it-IT`: Italian `kk-KZ`: Kazakh `nl-NL`: Dutch `pl-PL`: Polish `pt-PT`: Portuguese `pt-BR`: Brazilian Portuguese `sv-SE`: Swedish `tr-TR`: Turkish `uz-UZ`: Uzbek (Latin script)
`model`	`string`	Yes	No	No	Speech recognition model. Available model: `general` (general recognition model).

STT.ObjectStorage

Field name	Type	Required	Default value	Templating supported	Description
`bucket`	`string`	Yes	No	No	Name of the Object Storage bucket the audio file is in.
`object`	`string`	Yes	No	Yes	Object name, e.g., `prefix/subprefix/data.mp3`.

STT.Disk

Field name	Type	Required	Default value	Templating supported	Description
`oauthToken`	`string`	Yes	No	No	OAuth token for authentication when accessing Yandex Disk.
`path`	`string`	Yes	No	Yes	Audio file path.
`sharedDiskId`	`string`	No	No	Yes	Shared disk label.

Was the article helpful?

Previous

Next