Skip to content

Latest commit

 

History

History
33 lines (26 loc) · 2.15 KB

File metadata and controls

33 lines (26 loc) · 2.15 KB

AI Voice Connector - Community Edition - Deepgram Native Flavor

The Deepgram Native flavor uses the Deepgram's new Voice Agent API that provides direct Speech-to-Speech interpretation of the user's conversation. This setup allows for real-time interpretation and response generation, streamlining interactions without intermediate steps.

Implementation

The project uses the native WebSocket connection to push the decapsulated RTP from the user to the Deepgram engine and get the response back. Then, it grabs the response, packs it back by adding the RTP header, and streams it back to the user.

It does not have any transcoding capabilities, thus communication is limited to g711 PCMU and PCMA codecs.

Configuration

The following parameters can be tuned for this engine:

Section Parameter Environment Mandatory Description Default
deepgram_native key DEEPGRAM_API_KEY yes Deepgram API key not provided
deepgram_native speech_model DEEPGRAM_NATIVE_SPEECH_MODEL no Deepgram's speech detection model nova-2-conversationalai
deepgram_native voice DEEPGRAM_NATIVE_VOICE no Deepgram's voice used for speaking back the response aura-asteria-en
deepgram_native welcome_message DEEPGRAM_NATIVE_WELCOME_MSG no A welcome message to be played back to the user when the call starts ``
deepgram_native instructions DEEPGRAM_INSTRUCTIONS no Configures the LLM instructions ``
deepgram_native llm_url DEEPGRAM_LLM_URL no Configures the LLM URL ``
deepgram_native llm_auth DEEPGRAM_LLM_KEY no Configures the LLM API key ``
deepgram_native llm_model DEEPGRAM_LLM_MODEL no Configures the LLM model ``
deepgram_native disable DEEPGRAM_NATIVE_DISABLE no Disables the flavor false