Returns this STT's capabilities
Resolves once the scheduled fake_user_speeches have all been emitted.
Get the model name/identifier for this STT instance.
The model name if available, "unknown" otherwise.
Plugins should override this property to provide their model information.
Get the provider name for this STT instance.
The provider name if available, "unknown" otherwise.
Plugins should override this property to provide their provider information.
Channel: one sentinel per recognize() invocation.
Channel: one stream instance per stream() invocation.
Receives an audio buffer and returns transcription in the form of a SpeechEvent
Optional abortSignal: AbortSignalReturns a SpeechStream that can be used to push audio frames and receive transcriptions
Optional options: { Optional configuration including connection options
Optional connReplace one or more fake knobs mid-test (e.g. flip from error to success).
Configurable stand-in for a real STT. Knobs mirror Python's
FakeSTT: inject exceptions, scripted transcripts, connection timeouts, or a full sequence of FakeUserSpeech turns.Observability: every call to
recognize()posts to recognizeCh and every call tostream()posts the new stream to streamCh, so tests can assert on attempt counts directly instead of inferring them.Example