TranscribeAudioConfig class final

Configuration for the transcribeAudioTool transcription endpoint.

Constructors

TranscribeAudioConfig({String modelId = 'whisper-1', required String apiKey, String? baseUrl, String? language, Client? httpClient, TranscriptionResponseFormat responseFormat = TranscriptionResponseFormat.json, Set<TranscriptionGranularity> granularities = const {}, TranscriptionRequirement require = TranscriptionRequirement.none, String? prompt, TranscriptionOutputMode output = TranscriptionOutputMode.inline, int maxInlineBytes = defaultTranscribeMaxInlineBytes, TranscriptionProviderKind? providerKind})
Creates a configuration.
const

Properties

apiKey → String
API key for the transcription provider.
final
baseUrl → String?
Optional base URL of the Whisper-compatible endpoint; /audio/transcriptions is appended. When omitted, OpenAI's default is used (https://api.openai.com/v1). Changing this sends the audio (and the prompt) to that host — a configuration-trust decision.
final
granularities → Set<TranscriptionGranularity>
Requested timestamp granularities (only meaningful with verbose).
final
hashCode → int
The hash code for this object.
no setterinherited
httpClient → Client?
Optional HTTP client for testing.
final
language → String?
Optional ISO-639-1 language hint sent as the language form field.
final
maxInlineBytes → int
Inline byte budget for auto output mode.
final
modelId → String
Transcription model id, e.g. whisper-1 (OpenAI) or whisper-large-v3 (Groq). Defaults to whisper-1.
final
output → TranscriptionOutputMode
Result routing for verbose documents (see TranscriptionOutputMode).
final
prompt → String?
Optional domain-vocabulary hint (names, jargon, lyrics) forwarded verbatim to the endpoint; materially improves proper-noun fidelity.
final
providerKind → TranscriptionProviderKind?
Provider adapter family; inferred from the host when null.
final
require → TranscriptionRequirement
Minimum required timing level; downgrades below it are hard errors.
final
responseFormat → TranscriptionResponseFormat
Wire response format. json (default) keeps today's plain-text path byte-identically; verbose returns a TranscriptDocument; text returns the bare transcript body. Requesting granularities or a requirement implicitly upgrades json to verbose.
final
runtimeType → Type
A representation of the runtime type of the object.
no setterinherited
wantsDocument → bool
Whether the effective wire format should be verbose_json.
no setter

Methods

noSuchMethod(Invocation invocation) → dynamic
Invoked when a nonexistent method or property is accessed.
inherited
toString() → String
A string representation of this object.
inherited

Operators

operator ==(Object other) → bool
The equality operator.
inherited