Dynamically render react-components streamed from an LLM

[rajeshdavidbabu]

Hey Team,

What am I doing ?

I am currently building an internal tool for my company where I would dynamically render React components generated using Chatgpt using React Server Components.

The example I am using

I am using this official ai-sdk server component example, https://sdk.vercel.ai/docs/guides/frameworks/nextjs-app#in-server-components

What do I want ?

I want something similar, but instead of rendering the text, I need to parse and render React component form the stream data and return them.

Kinda similar to V0.dev's approach when we run the initial render.

[Kaidanov]

did you fined an answer?

[rajeshdavidbabu]

Currently you can do this already with AI SDK. Checkout https://www.morphic.sh/ and its open-source code-base. https://www.youtube.com/watch?v=XdaiGK3ySj0&t

[dnascimento]

Looking for the same but with Vite