Build a Multi-LLM Voice Assistant in 12 Minutes with Next.JS

[@stonedizzleful]

Fantastic video. So many features in one package!!

[@DevelopersDigest]

Thank you Liam! 🙏

[@mjohnson9104]

Wow…this is goal setting.

[@hellagood9]

Thanks!!!

[@DevelopersDigest]

Thanks for watching!

[@Amplify-AI]

For conversational memory, would you recommend storing that in the client state and using context stuffing when sending the prompt?

[@Amplify-AI]

I was thinking of using the OpenAI agents API which supports threads but each vendors SDK would be very different and hard to unify. So perhaps it’s just easier to use the Vercel AI SDK

[@DevelopersDigest]

@@Amplify-AI Good question - the nice thing with the vercel ai sdk is that you can set up the continuity between state of the chat between the front and back end and it is vendor agnostic. Langchain also has a nice template if you are looking for something in next.js as well.
github.com/langchain-ai/langchain-nextjs-template/

[@phinneasmctipper7518]

This is so cool!

[@DevelopersDigest]

Thanks Phinneas!!

[@pineappleeeeeeee2920]

Hi, can we integrate OpenAI TTS / deepgram here instead of webkit?

[@prashlovessamosa]

Thank you.

[@DevelopersDigest]

Thank you for watching 🙏

[@Amplify-AI]

What an amazing demo and valuable collection of resources. Thank you for sharing this.
Would this be possible to run using Google VertexAI sdk for speech to text, text to speech, and multimodal vision capabilities?

[@DevelopersDigest]

AI Device: github.com/developersdigest/ai-devices
Gemini Flash Multimodal example: github.com/developersdigest/gemini-flash-api

[@mrquicky]

The remote execution of LLM's through a licensed third party in the clouds. Saves electricity and hardware costs at home!

[@DevelopersDigest]

Well put! 🙂 Thanks for watching!