barreiros/deepgram-groq-realtime-voice-interaction

Normalize the sample audio rate for the capture and playback

Fix Vite

Configure npm install for production

Add the package-lock for production npm ci command run

Reduce the amount of data saved in memory

Improve sbert utterance detection

First version of SBERT working in local whit Docker

Improving the sentence detector

Refactoring layour WIP

Refactoring the frontend

Cleaning

Hangle multiple images

Images description working well

Adding and hangling multiple images

First version of the add images functionality

Fix onMessage callback

Configure fallback url

Add env variables

Add secret prompt

Modify npm in Dockerfile

Rename Readme.md to README.md

Add headers

Add headers

Add headers

Add headers

Configuring Dockerfile for Hugginface spaces

First version of the Dockerfile

Add encoding parameter

Prevent some errors

Hangle Groq API errors

Set defaults

Separate the params by service

Improving shutup system

Stop the current conversation when user talk

Configure the shutup process using an LLM when sentences are not completed

feat: update GroqService model and conversation chain input

fix: clear listen connection buffers in DeepgramService clearBuffers method

feat: add clearBuffers to GroqService and shutup event handling

feat: add timer functionality to send messages at intervals

Stop conversation WIP

Stop conversation WIP

Stop conversation WIP

Stop conversation WIP

Stop conversation WIP

First version of stop conversation working

Init the shutup implementation

refactor: remove unused state variable inputMessage

Refactor some classes

add handling for flushing word buffer when playback queue is empty

feat: update Deepgram TTS model to 'nova-3' and handle speak flush event