Update README.md

feat: landing page

replaced nextjs monorepo with vite + react (best decision ever)

update: readme.md

heavy commit: next js monorepo ai slop which works with the current backend state

blueprint for realtime transcription

json based redis storage

add: initial user input at runtime via create session

changed iteration to counter based

user input as text now on frontend

Implement audio playback state management and websocket events

streaming pcm data and resolving on frontend

stream bytes: pocket tts

asynchronous wrapper for pocket tts

init pocket tts