Usage guide
Setup, environment variables, and Python/JavaScript/Go usage examples.
GladiaSTT
Real-time speech-to-text with Gladia, using your own Gladia key. Strong on multilingual conversations, with automatic switching between languages mid-call. Authenticates with GLADIA_API_KEY from your environment, or an explicitapi_key.
Fields
str
default:"UNSET"
Gladia transcription model, e.g. “solaria-1”. Defaults to ‘solaria-1’.
List[str] | None
default:"UNSET"
Expected languages as ISO 639-1 codes, e.g. [“en”, “fr”]. Leave unset to let Gladia detect the language. Defaults to None (not set).
bool
default:"UNSET"
Whether the speaker may switch languages mid-conversation, with transcription following along. Defaults to True.
int | None
default:"UNSET"
Sample rate of incoming audio in Hz. Leave unset to detect it automatically. Defaults to None (not set).
int | None
default:"UNSET"
Sample rate in Hz sent to Gladia: 8000, 16000, 32000, 44100 or 48000. Leave unset to use the incoming rate, or the nearest supported one. Defaults to None (not set).
str
default:"UNSET"
Encoding of the audio sent to Gladia, e.g. “wav/pcm”, “wav/alaw” or “wav/ulaw”. Defaults to ‘wav/pcm’.
int
default:"UNSET"
Bits per sample of the audio sent to Gladia, e.g. 8, 16, 24 or 32. Defaults to 16.
int | None
default:"UNSET"
Number of audio channels the agent sends to Gladia. Leave unset for mono; Gladia bills per channel. Defaults to None (not set).
bool
default:"True"
Whether Gladia sends partial transcripts while the user is still speaking. Defaults to True.
bool
default:"True"
Whether partial transcripts are also sent to the client alongside final ones. Turn detection uses partials either way. Defaults to True.