While building Nuxt, we created a new server engine: Nitro.
It is shipped with many features:
Server API endpoints and Middleware are added by Nitro that internally uses h3.
Key features include:
res.end() and next() are also supported)Check out the h3 docs for more information.
Nitro allows 'direct' calling of routes via the globally-available $fetch helper. This will make an API call to the server if run on the browser, but will directly call the relevant function if run on the server, saving an additional API call.
$fetch API is using ofetch, with key features including:
Content-Type headersFor more information on $fetch features, check out ofetch.
When using API routes (or middleware), Nitro will generate typings for these routes as long as you are returning a value instead of using res.end() to send a response.
You can access these types when using $fetch() or useFetch().
Nitro produces a standalone server dist that is independent of node_modules.
The server in Nuxt 2 is not standalone and requires part of Nuxt core to be involved by running nuxt start (with the nuxt-start or nuxt distributions) or custom programmatic usage, which is fragile and prone to breakage and not suitable for serverless and service worker environments.
Nuxt generates this dist when running nuxt build into a .output directory.
The output contains runtime code to run your Nuxt server in any environment (including experimental browser service workers!) and serve your static files, making it a true hybrid framework for the JAMstack. In addition, Nuxt implements a native storage layer, supporting multi-source drivers and local assets.