Yandex Speech STTInstall

Summary

Transcribes an MP3, WAV, or OGG audio file with Yandex SpeechKit using its asynchronous file recognition API. Provide audio as binary data from the previous node or as an object in Yandex Object Storage; the operation waits for recognition to finish before returning results.

Inputs

• fileSource (required) — Choose whether to read the audio file from a previous node's binary data or from Yandex Object Storage. Defaults to the previous node.
• binaryFieldName (required) — When using the previous node, name the binary field containing the audio file. Defaults to data; the binary data must also include a filename.
• accessKeyId (required) — For Yandex Object Storage, the AWS-compatible access key ID.
• secretAccessKey (required) — For Yandex Object Storage, the AWS-compatible secret access key.
• s3Region (required) — For Yandex Object Storage, the bucket region. Defaults to ru-central1.
• s3Endpoint (required) — For Yandex Object Storage, the storage endpoint. Defaults to https://storage.yandexcloud.net.
• bucketName (required) — For Yandex Object Storage, the bucket containing the audio file.
• objectKey (required) — For Yandex Object Storage, the object's key, including the audio filename; the key must end in .mp3, .wav, or .ogg.

Output shape

One item containing json.yandexSTTRecognition, an object with text and rawResult fields.

text concatenates the text from final recognition alternatives; rawResult contains each parsed recognition result returned by the service. The node uses its required Yandex Speech API credential (apiKey and folderId) in addition to the selected file source. It requests the general recognition model, polls until the operation completes, and then fetches the results. For Object Storage input, it creates a presigned object URL that expires after 100 seconds. Invalid or missing audio data and service/API errors fail the operation.

Links

• n8n community nodes installation guide