Skip to content

DeepgramTranscriptionOptions

Deepgram-specific transcription options passed as query parameters on the WebSocket connection URL.

Defined in: src/providers/stt/deepgram/DeepgramSTT.ts:28

Deepgram-specific transcription options passed as query parameters on the WebSocket connection URL.

Remarks

These options map to Deepgram’s real-time streaming API query parameters. They are set on the DeepgramSTTConfig.options property and appended to the WebSocket URL when the connection is established.

See

DeepgramSTTConfig for the full provider configuration

Properties

PropertyTypeDefault valueDescriptionDefined in
alternatives?numberundefinedNumber of transcription alternatives to return (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:103
channels?numberundefinedNumber of audio channels.src/providers/stt/deepgram/DeepgramSTT.ts:55
detectEntities?booleanundefinedExtract named entities from the transcript (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:123
diarize?booleanundefinedEnable diarization / speaker detection (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:88
dictation?booleanundefinedEnable dictation mode (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:132
encoding?stringundefinedAudio encoding format. Supported: 'linear16', 'linear32', 'flac', 'alaw', 'mulaw', 'amr-nb', 'amr-wb', 'opus', 'ogg-opus', 'speex', 'g729'.src/providers/stt/deepgram/DeepgramSTT.ts:49
endpointing?number | boolean10Automatic endpointing — milliseconds of silence before end-of-speech. Set to false to disable (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:117
extra?string[]undefinedArbitrary key:value metadata passed through to the API (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:161
interimResults?booleanundefinedEnable interim results (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:109
keyterms?string[]undefinedSpecialized terminology to boost recognition. Unlike keywords, keyterms use a more advanced boosting algorithm. Example ['CompositeVoice', 'WebSocket']src/providers/stt/deepgram/DeepgramSTT.ts:64
keywords?string[]undefinedCustom vocabulary or keywords to boost or suppress recognition (V1 only). Append a weight to boost/suppress: e.g. 'Deepgram:2' or 'erm:-10'. See keyterms for the newer boosting API (V1 & V2).src/providers/stt/deepgram/DeepgramSTT.ts:100
language?stringundefinedLanguage code (e.g., 'en-US', 'es').src/providers/stt/deepgram/DeepgramSTT.ts:41
mipOptOut?booleanundefinedOpt out of the Deepgram Model Improvement Program.src/providers/stt/deepgram/DeepgramSTT.ts:70
model?stringundefinedNova model to use for V1 transcription. Available models: 'nova-3' (default), 'nova-3-medical', 'nova-2', 'nova-2-general', 'nova-2-meeting', 'nova-2-finance', 'nova-2-conversationalai', 'nova-2-voicemail', 'nova-2-video', 'nova-2-medical', 'nova-2-drivethru', 'nova-2-automotive', 'nova', 'enhanced', 'base'.src/providers/stt/deepgram/DeepgramSTT.ts:38
multichannel?booleanundefinedTranscribe each audio channel independently (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:129
numerals?booleanundefinedConvert spoken numbers to digit form, e.g. “twenty one” → “21” (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:126
profanityFilter?booleanundefinedEnable profanity filter (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:78
punctuation?booleanundefinedEnable punctuation (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:75
redact?string[]undefinedEnable redaction of sensitive information (V1 only). Example ['pci', 'ssn', 'numbers']src/providers/stt/deepgram/DeepgramSTT.ts:85
replace?string[]undefinedTerm replacement — swap recognized terms for preferred forms (V1 only). Example ['colour:color', 'grey:gray']src/providers/stt/deepgram/DeepgramSTT.ts:139
sampleRate?numberundefinedSample rate for audio data in Hz (required when encoding is set).src/providers/stt/deepgram/DeepgramSTT.ts:52
search?string[]undefinedSearch for specific terms or phrases in the transcript (V1 only). Example ['action item', 'follow up']src/providers/stt/deepgram/DeepgramSTT.ts:146
smartFormat?booleanundefinedEnable smart formatting — auto-punctuation and readability improvements (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:91
tag?stringundefinedLabel for usage reporting in the Deepgram console.src/providers/stt/deepgram/DeepgramSTT.ts:67
utteranceEndMs?numberundefinedDelay in milliseconds before emitting an UtteranceEnd event (V1 only). Controls the gap required between utterances. Example 1000src/providers/stt/deepgram/DeepgramSTT.ts:155
utterances?booleanundefinedEnable utterance segmentation (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:106
vadEvents?booleanundefinedEmit Voice Activity Detection events (SpeechStarted) (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:120
version?stringundefinedModel version override (V1 only).src/providers/stt/deepgram/DeepgramSTT.ts:158

© 2026 CompositeVoice. All rights reserved.

Font size
Contrast
Motion
Transparency