tRPC server documentation

📄️ Define Routers

To begin building your tRPC-based API, you'll first need to define your router. Once you've mastered the fundamentals, you can customize your routers for more advanced use cases.

📄️ Define Procedures

- A publicProcedure can be viewed as the equivalent of a REST-endpoint or a function.

📄️ Merging Routers

Writing all API-code in your code in the same file is not a great idea. It's easy to merge routers with other routers.

📄️ Request Context

The createContext() function is called for each request, and the result is propagated to all resolvers. You can use this to pass contextual data down to the resolvers.

📄️ Middlewares

You are able to add middleware(s) to a procedure with the t.procedure.use() method. The middleware(s) will wrap the invocation of the procedure and must pass through its return value.

📄️ Server Side Calls

You may need to call your procedure(s) directly from the server, createCaller() function returns you an instance of RouterCaller able to execute queries and mutations.

📄️ Authorization

The createContext function is called for each incoming request, so here you can add contextual information about the calling user from the request object.

tRPC gives you automatic type-safety of outputs without the need of adding a validator; however, it can be useful at times to strictly define the output type in order to prevent sensitive data from being leaked.

📄️ Inferring Types

It is often useful to wrap functionality of your @trpc/client or @trpc/react-query api within other functions. For this purpose, it's necessary to be able to infer input types and output types generated by your @trpc/server router.

📄️ Error Handling

Whenever an error occurs in a procedure, tRPC responds to the client with an object that includes an "error" property. This property contains all the information that you need to handle the error in the client.