STT Connector

STT Connector

Speech-to-Text Setup Options.

This page edits the single global STT connector. Switching services updates the active runtime provider instead of creating extra connector records.
Recommended
Deepgram
DEEPGRAM
Deepgram's Whisper Speech-to-Text
Parakeet
PARAKEET
parakeet-api-server
Other Services
Whisper
WHISPER
OpenAI's Whisper
Local Whisper
LOCALWHISPER
Local Whisper (Installed in DwemerDistro)
Gemini
GEMINI
Google Gemini STT + Emotion Detection
Azure
AZURE
Azure Speech-to-Text
Inworld
INWORLD
Inworld STT
System
Disabled
SYSTEM
Disabled
Testing saves the current connector first so the modal uses the latest settings.
This label is kept for migration and internal reference, even though only one STT connector is used globally.
Choose the speech-to-text backend HerikaServer should load globally.
Cloud STT services require an API key from the API Keys page.

Deepgram's Whisper Speech-to-Text Settings

Language
Model to use

parakeet-api-server Settings

Language to detect for STT.

OpenAI's Whisper Settings

Language to detect for STT.
Will try to translate to english.

Local Whisper (Installed in DwemerDistro) Settings

Form field name for audio file. Sometimes needed to change to file to use another shiper implementations

Google Gemini STT + Emotion Detection Settings

Language (e.g., en, es, ja)
Gemini model. 2.5 Flash recommended for best speed and quality.

Azure Speech-to-Text Settings

Language to detect for STT.
Specifies how to handle profanity in recognition results. Accepted values are:
MASKED, which replaces profanity with asterisks.
REMOVED, which removes all profanity from the result.
RAW, which includes profanity in the result.

Inworld STT Settings

Model identifier in provider/model format, e.g. groq/whisper-large-v3.
Language code in BCP-47 format, e.g. en-US. Leave blank for auto-detect.

Disabled Settings

This STT provider does not have connector-level settings to configure here.