API v2 for streaming recognition

The streaming recognition service is located at stt.api.cloud.yandex.net:443.

Message with recognition settings

Parameter	Description
config	object Field with the recognition settings and folder ID.
config .specification	object Recognition settings.
config .specification .languageCode	string Recognition language. See the model description for valid values. The default value is `ru-RU`, Russian.
config .specification .model	string Language model to use for recognition. The more accurate your choice of the model, the better the recognition result. You can only specify one model per request. Valid values depend on the selected language. The default value is `general`.
config .specification .profanityFilter	boolean Profanity filter. Valid values: `true`: Exclude profanity from the recognition results. `false` (default): Do not exclude profanity from the recognition results.
config .specification .partialResults	boolean Intermediate result filter. Acceptable values: `true`: Return intermediate results, i.e., a part of the recognized utterance. For intermediate results, `final` equals `false`. `false` (default): Only return the final results, i.e., the entire recognized utterance.
config .specification .singleUtterance	boolean Flag disabling recognition after the first utterance. Acceptable values: `true`: Recognize only the first utterance, stop recognition and wait for the user to terminate the session. `false` (default): Continue recognition until the end of the session.
config .specification .audioEncoding	string Audio format. Valid values: `LINEAR16_PCM`: LPCM without a WAV header. `OGG_OPUS` (default): OggOpus format.
config .specification .sampleRateHertz	integer (int64) Audio sampling rate. This parameter is required if `format` is set to `LINEAR16_PCM`. Valid values: `48000` (default): 48 kHz. `16000`: 16 kHz. `8000`: Sampling rate of 8 kHz.
config. specification. rawResults	boolean Flag that toggles spelling out numbers: `true` for words, `false` (default) for figures.
folderId	string ID of the folder you have access to. It is required for authentication with a user account (see Authentication with the SpeechKit API). Do not use this field if you make a request on behalf of a service account. The maximum string length is 50 characters.

Experimental additional recognition settings

Streaming recognition models support new recognition settings. They are provided in metadata in a gRPC call.

Parameter	Description
`x-normalize-partials`	boolean Flag allowing you to get intermediate recognition results (parts of the recognized utterance) in a normalized format, i.e., numbers presented as digits, profanity filter enabled, etc. Valid values: `true`: Return a normalized result. `false` (default): Return a non-normalized result.

Audio message

Parameter	Description
`audio_content`	Audio fragment as an array of bytes. The audio format must match the value specified in the message with recognition settings.

Message with recognition results

If the speech fragment recognition is successful, you will receive a message containing a list of recognition results (chunks[]). Each result contains the following fields:

alternatives[]: List of recognized text alternatives. Each alternative contains the following fields:
- text: Recognized text.
- confidence: This field is currently not supported. Do not use it.
final: Flag indicating that this recognition result is final and will not change anymore. If the value is false, it means the recognition result is intermediate and may change as subsequent speech fragments get recognized.
endOfUtterance: Flag indicating that this result contains the end of the utterance. If the value is true, the new utterance will start with the next result you get.

Note

If you set singleUtterance=true, only one utterance per session will be recognized. After the message where endOfUtterance is true, the server will not recognize any following utterances and will wait for you to terminate the session.

Error codes returned by the server

To see how gRPC statuses map to HTTP codes, see google.rpc.Code.

List of possible gRPC errors returned by the service:

Code	Status	Description
3	`INVALID_ARGUMENT`	Invalid request parameters. Detailed information is provided in the `details` field.
9	`RESOURCE_EXHAUSTED`	Client exceeded a quota.
16	`UNAUTHENTICATED`	The operation requires authentication. Check the IAM token and the folder ID that you provided.
13	`INTERNAL`	Internal server error. This error indicates that the operation cannot be completed due to a server-side issue, e.g., insufficient computing resources.