Skip to content
A simple example of creating pages dynamically in Gatsby without using GraphQL.
Branch: master
Clone or download

Can you use Gatsby without GraphQL?


This is a small example that loads data from the PokéAPI’s REST endpoints, then creates pages (and nested pages) using Gatsby’s createPages API.

What are the trade-offs?

The upsides of using REST:

  • The approach is familiar and comfortable, especially if you’re new to GraphQL
  • There’s no intermediate step: you fetch some data, then build pages with it

The downsides of using REST:

  • There are lots of calls required, and each nested call relies on data from the previous call; this won’t scale well
  • All of the data for the page needs to be explicitly passed into the context object, which makes it a little harder to understand what data is being passed to the page component
  • The relationships between items are harder to understand; we need to make three separate requests, resulting in three separate data objects that we have to manually combine in the front-end

What would this look like querying from a GraphQL API?

Great question! There’s not a stable Pokémon GraphQL API that I’ve seen, but if there was, the query might look like this:

const data = await graphql(`
    Pokemon {
      edges {
        node {
          abilities {

// Use createPage to turn the data into pages, just like the REST version.

This one query accomplishes the same thing as the three different REST calls, and it shows more clearly how the data is related (e.g. each Pokémon has abilities).

What would this look like using Gatsby's GraphQL integration layer?

For quick and easy comparison, the using-gatsby-data-layer branch illustrates how you can accomplish this using Gatsby's integration layer, rather than using the unstructured data approach.

You can’t perform that action at this time.