Documentation
¶
Overview ¶
Package dictation provides an embeddable strict Dictation runtime.
Dictation is the most boundary-strict of the three SpeechKit modes: STT turns speech into text and the pipeline stops there. No LLM rewriting, no codewords, no tool invocation. Host apps integrate this when they want a transcript surface they can route into their own editor, command dispatcher, or downstream processor.
Use NewRuntime to construct an instance; pass a recorder, a worker, and the speechkit.RuntimePolicy from the host config.
Index ¶
Constants ¶
This section is empty.
Variables ¶
View Source
var ( ErrMissingRecorder = errors.New("speechkit dictation: recorder is required") ErrMissingTranscriber = errors.New("speechkit dictation: transcriber is required") ErrAlreadyRecording = errors.New("speechkit dictation: already recording") ErrNotRecording = errors.New("speechkit dictation: not recording") ErrAudioTooShort = errors.New("speechkit dictation: audio too short") )
Functions ¶
This section is empty.
Types ¶
type Options ¶
type Options struct {
Recorder speechkit.AudioRecorder
Transcriber speechkit.Transcriber
Output speechkit.TranscriptOutput
Store speechkit.Persistence
Policy speechkit.RuntimePolicy
Profiles []speechkit.ProviderProfile
Language string
Target any
MinPCMBytes int
}
type Runtime ¶
type Runtime struct {
// contains filtered or unexported fields
}
Runtime is an embeddable Dictation-only service. It keeps the mode boundary strict: audio in, final text out, no tool calls or LLM rewriting.
func NewRuntime ¶
Source Files
Click to show internal directories.
Click to hide internal directories.