Improve frontend performance (+ integrate sync engine)

[tobiasbueschel]

@t3dotgg and his team have done a great job with https://t3.chat especially when it comes to performance. It would be interesting to explore what concepts we could incorporate here in the ai-chatbot template to fully optimize and showcase the performance of Next.js.

Here is an overview of how T3 chat was built:

• https://www.youtube.com/watch?v=QLvIoi2s1zY

Sync engines could be interesting to explore

• https://developer.mozilla.org/en-US/docs/Web/API/IndexedDB_API
• https://github.com/dexie/Dexie.js
• https://news.ycombinator.com/item?id=43397640
• https://www.instantdb.com/essays/sync_future
• https://dev.to/isaachagoel/are-sync-engines-the-future-of-web-applications-1bbi
• https://zero.rocicorp.dev
• https://jazz.tools

React Scan / Million

• Do we want to integrate something like: https://github.com/aidenybai/react-scan or https://million.dev/ to check if any rendering could be improved?

Layout.tsx prefetching

• Also, this video from @leerob explains the concept of prefetching requests on the server-side using layout.tsx, which could be interesting to explore: https://x.com/leerob/status/1901349669554290826

Caching using AI SDK

• https://sdk.vercel.ai/cookbook/next/caching-middleware

---

Has anyone looked into any of the above already using the ai-chatbot template?