GoogleTranscriptionModel

Namespace: Vercel.AI.Google

public sealed class GoogleTranscriptionModel : ITranscriptionModel, ITranscriptionCaller

Gemini transcription. Unary models post to the Interactions API.

Inheritance: object

Implements: ITranscriptionModel, ITranscriptionCaller

View source

Constructors

GoogleTranscriptionModel(GoogleProvider, string)

public GoogleTranscriptionModel(GoogleProvider provider, string modelId)

Creates a transcription model.

Parameters

Name	Type	Description
provider	GoogleProvider
modelId	string

View source

Properties

CanStream

public bool CanStream { get; }

False. The Gemini Live WebSocket is not opened by this port.

Value: bool

View source

ModelId

public string ModelId { get; }

Model id.

Value: string

View source

Provider

public string Provider { get; }

Provider id.

Value: string

View source

SpecificationVersion

public string SpecificationVersion { get; }

Specification version.

Value: string

View source

Methods

BuildLiveSetup(string)

public static JsonObject BuildLiveSetup(string modelId)

Builds the Live API setup message. Live models reject a unary call.

Parameters

Name	Type	Description
modelId	string

Returns: JsonObject

View source

BuildRequest(string, AudioInput, JsonObject?)

public static JsonObject BuildRequest(string modelId, AudioInput audio, JsonObject? transcriptionConfig)

Builds the Interactions transcription body.

Parameters

Name	Type	Description
modelId	string
audio	AudioInput
transcriptionConfig	JsonObject

Returns: JsonObject

View source

DoGenerateAsync(TranscriptionModelCall, CancellationToken)

public Task<TranscriptionModelResult> DoGenerateAsync(TranscriptionModelCall call, CancellationToken cancellationToken)

Transcribes audio through the Interactions API. google provider options become the transcription_config, and word_info annotations become word segments.

Parameters

Name	Type	Description
call	TranscriptionModelCall
cancellationToken	CancellationToken

Returns: Task<TranscriptionModelResult>

View source

DoStreamAsync(TranscriptionStreamCall, CancellationToken)

public Task<TranscriptionStreamStart?> DoStreamAsync(TranscriptionStreamCall call, CancellationToken cancellationToken)

Returns null. The Gemini Live WebSocket is not opened by this port.

Parameters

Name	Type	Description
call	TranscriptionStreamCall
cancellationToken	CancellationToken

Returns: Task<TranscriptionStreamStart>

View source

DoTranscribeAsync(AudioInput, CancellationToken)

public Task<TranscriptionResult> DoTranscribeAsync(AudioInput audio, CancellationToken cancellationToken)

Transcribes audio.

Parameters

Name	Type	Description
audio	AudioInput
cancellationToken	CancellationToken

Returns: Task<TranscriptionResult>

View source

IsLive(string)

public static bool IsLive(string modelId)

True when the model id is a Live API transcription model.

Parameters

Name	Type	Description
modelId	string

Returns: bool

View source