t3 monorepo with fastify backend
Greetings!
I’ve enjoyed building with the t3 stack for over a year and scaled a platform to production without any need for a server.
Now I am running against the limits with vercels function timeouts, e.g. when using transcription APIs that need minutes to respond, complex llm calls with langchain etc.
I am currently looking for a good solution to use nextjs with tRPC + RSCs + a dedicated fastify backend for long running tasks in a monorepo.
With the new paradigm of RSC I am wondering where to draw the line between responsibilities - should both nextjs and fastify use tRPC and prisma? Is it better to have all business logic on fastify instead of splitting it or to only use it for the more complex tasks?
I am curious if anyone has made up their mind about this already or if anyone can recommend a good template or starter for this setup.
Thank you!
I’ve enjoyed building with the t3 stack for over a year and scaled a platform to production without any need for a server.
Now I am running against the limits with vercels function timeouts, e.g. when using transcription APIs that need minutes to respond, complex llm calls with langchain etc.
I am currently looking for a good solution to use nextjs with tRPC + RSCs + a dedicated fastify backend for long running tasks in a monorepo.
With the new paradigm of RSC I am wondering where to draw the line between responsibilities - should both nextjs and fastify use tRPC and prisma? Is it better to have all business logic on fastify instead of splitting it or to only use it for the more complex tasks?
I am curious if anyone has made up their mind about this already or if anyone can recommend a good template or starter for this setup.
Thank you!
