make voice livekit work livekit skills voice/debo-ai talk and recording voice ui are different keep them diff pages. also find the gaps if you need anything else to be added do it. and make a proper plan and deep dive on this. and make the voice working.
create the best artitecture for http://localhost:3000/dashboard/brain and http://localhost:3000/dashboard/twin also suggest which cf services I can also include in this project also create a plan to ui ux revamp minimal professional , revamp design and ui ux of http://localhost:3000/dashboard/voice/talk and all pages
fetch details about use https://agent-elements.21st.dev/ use context7 and seaches and create plan to use it and replce all existing things
allow digital twin sharing feature http://localhost:3000/dashboard/twin make ui more better http://localhost:3000/dashboard/brain
in all pages in the sidebar and the whole app i.e. = http://localhost:3000/dashboard/mcp and all remove all custom css and custom components I want you to use shadcn skills and use pure shadcn components without modification directly raw everywhere and replace all custom components maximum use the tokens defined in the globalscss of shadcn defaults and revamp the clean minimal ui with raw shadcn and keep things afster from frontend