Integrating Contentlayer with NextJs Blog

Blog Setup Development NextJs Contentlayer

The second part of this series explores Contentlayer and how to integrate it with a NextJs blog. I will elaborate on my experience managing content on my blog throughout the last eight years using various tools, why I decided to use Contentlayer and how to integrate it into a NextJs blog.

History

When I was starting out on my journey with NextJs to build the next version of my blog, I had to figure out a way to feed data into my application. I used only traditional static site generators like Jekyll and Hugo until then. SSGs work in this way - One writes the content in local markdown files; the SSG takes these files, extract the front matter and actual body, converts the extracted information to variables and feeds them into the actual server-side rendering logic, which will generate the static HTML files. Storing these content markdown files in Git is what is today called the Git CMS. I am a big proponent of keeping the content in a Git repository because it makes it easier to track the writing journey without tying too much into a vendor.

Although the concept of SSGs came into being with the release of Jekyll in 2008 (at the same time as Github and its iconic Pages feature), I actually started my journey with it in 2015. Since then, technology has advanced by leaps and frogs, and several new approaches to managing content and code have emerged.

I also dabbled with integrating headless CMS (WordPress , Sanity and Contentful ) and GraphQL based Gatsby at various points in past few years, however I was never successful because of my unease with JavaScript. My unwillingness to rely on any third-party vendor for storing my content and the costs associated with self-hosting any open source CMS also contributed greatly to my failure to transition.

Migrating to NextJs

When I was starting to migrate my blog to NextJs, I had a couple of requirements in mind -

I preferred using a headless CMS for storing the content. A few reasons are - a. My writing was already suffering because of my over-emphasis on coding instead of writing. b. I was somewhat frustrated with my blog's overly complicated publishing process.
The CMS should allow export in popular open formats when I decide to migrate in future.

Notion fits nicely into my requirements. It has a well-defined (although with poor performance) API . It is a popular tool with the developer community and offers the export of the content in markdown format. Notion would have worked well for my requirements, but I couldn't integrate it properly with my blog due to my lack of JavaScript experience.

So, I reverted back to using markdown for my blog's content. I still write in Notion's excellent editor but then export the file to markdown and manually edit image paths etc., to work with my blog's structure.

However, using md and its more advanced sibling mdx with NextJs is tedious and error-prone. A bunch of tools need to be integrated to make MDX work with NextJs. unified, remark and rehype ecosystem provides some great tools, but combining these is not trivial. I had already wasted enough time struggling with Notion. I might have abandoned the project altogether if I had wasted a few more days.

My search finally ended on Contentlayer . Contentlayer is a tool which transforms unstructured markdown content into structured type-safe JSON data structures and provides accessible interfaces to access these data structures. This means I could import all my content in the code files using an import statement and work with the data as if working with a Typescript interface. This eliminated the need to struggle with additional boilerplate code needed to use the low-level libraries to interact with the MDX files.

With that in mind, let's see how to integrate Contentlayer in our NextJs blog. In later posts of this series, we will also see how to utilise the extensive plugin ecosystem provided by the unified project to enrich the data interfaces exposed by Contentlayer.

Install and Configure Contentlayer

Contentlayer's getting started guide is a good resource to setup your NextJs project to work with Contentlayer. I have described the same steps below for quick reference -

First, add contentlayer and next-contentlayer to the project -

yarn add contentlayer next-contentlayer

The next-contentlayer provides an API to stitch together next and contentlayer by hooking into the next dev and next build commands. This will ensure that contentlayer runs automatically for local development and production deployments.

Now, we need to patch the next.config.js to use the API provided by the next-contentlayer to hook into the commands mentioned above -

// next.config.js
const { withContentlayer } = require("next-contentlayer")
 
module.exports = withContentlayer(nextConfig)

Content Schema

Once we have contentlayer setup, it is time to define the structure of our data that will be used by Contentlayer to parse the MDX files. Contentlayer can also process the parsed content to derive new fields. Let's see how it is done.

Create a new file, contentlayer.config.js, in the root of your project. Inside this file, I am defining a Post document type -

import { defineDocumentType, makeSource } from "contentlayer/source-files"
 
export const Post = defineDocumentType(() => ({
  // The name of the document type
  name: "Post",
 
  // The type of the files to parse. 'md' also works
  contentType: "mdx",
 
  // The path of the mdx files, relative to contentDirPath
  filePathPattern: "posts/*.mdx",
 
  // Fields present in the frontmatter of the MDX file
  fields: {
    title: { type: "string", required: true },
    published: { type: "string", required: true },
    description: { type: "string" },
    status: {
      type: "enum",
      options: ["draft", "published"],
      required: true,
    },
  },
}))
 
export default makeSource({
  contentDirPath: "content",
  documentTypes: [Post],
})

In the code snippet above, the fields key is used to store the custom fields related to the metadata of a post, such as title, publish date etc. Any field can be made mandatory by using required: true.

Now, create a folder called content in the project's root, and inside this folder, create a subfolder called posts. We will store the mdx files inside this folder. Let's start our first post, content/posts/hello-world.mdx.

---
title: Hello World
published: 2023-07-30T13:05:24.000Z
status: published
---
 
Hello world! This is my first post.

If next dev is running in the background while you save this file, you will see the following output in your terminal -

wait - compiling...
File updated: content/posts/hello-world.mdx
Generated 1 document in .contentlayer

Contentlayer Artifacts

When next dev is run for the first time after setting up Contentlayer, Contentlayer creates a hidden folder inside the project root directory - .contentlayer. All the processed files, generated data structures, and helper functions and interfaces are stored here. Let's see the contents of this directory -

❯ tree
./
├── .contentlayer/
│   ├── generated/
│   │   ├── Post/
│   │   │   ├── _index.json
│   │   │   ├── _index.mjs
│   │   │   └── posts__hello-world.mdx.json
│   │   ├── index.d.ts
│   │   ├── index.mjs
│   │   └── types.d.ts
│   └── package.json
├── content/
│   └── posts/
│       └── hello-world.mdx
├── contentlayer.config.js
├── next.config.js
└── tsconfig.json

Contentlayer converts the static MDX file (line 15) into a structured JSON file (line 8). All such JSON files are accessible through the allPosts variable exported from index.mjs on line 7. In addition, it also generates type definitions (line 11) which helps Typescript understand the interface provided by the allPosts import.

// .contentlayer/generated/Post/posts__hello-world.mdx.json
 
{
  "title": "Hello World",
  "published": "2023-07-30T13:05:24.000Z",
  "status": "published",
  "body": {
    "raw": "\nHello world! This is my first post.\n",
    "code": "var Component=(()=>{var sr=Object.create;..."
  },
  "_id": "posts/hello-world.mdx",
  "_raw": {
    "sourceFilePath": "posts/hello-world.mdx",
    "sourceFileName": "hello-world.mdx",
    "sourceFileDir": "posts",
    "contentType": "mdx",
    "flattenedPath": "posts/hello-world"
  },
  "type": "Post"
}

Below is the generated type definition for the Post document type -

// .contentlayer/generated/types.d.ts
 
/** Document types */
export type Post = {
  /** File path relative to `contentDirPath` */
  _id: string
  _raw: Local.RawDocumentData
  type: 'Post'
  title: string
  published: string
  description?: string | undefined
  status: 'draft' | 'published'
  /** MDX file body */
  body: MDX
}

Computed Fields

Once Contentlayer has generated the JSON files, the content is ready to consume on our pages. We will use NextJs's dynamic routes to render a single blog post. But before that, we must decide which unique value will be used for navigating the page. File path without the file extension is a standard approach used in the community (called a slug), but we currently don't have this value available easily.

That's where Contentlayer's Computed Fields come into the picture. It lets us compute new metadata from existing content. For example, it can be used to calculate the estimated reading time from the post content.

To compute the slug field, we need to modify the contentlayer.config.js as follows -

export const Post = defineDocumentType(() => ({
  // The name of the document type
  name: "Post",
 
  // The type of the files to parse. 'md' also works
  contentType: "mdx",
 
  // The path of the mdx files, relative to contentDirPath
  filePathPattern: "posts/*.mdx",
 
  // Fields present in the frontmatter of MDX file
  fields: {
    title: { type: "string", required: true },
    published: { type: "string", required: true },
    description: { type: "string" },
    status: {
      type: "enum",
      options: ["draft", "published"],
      required: true,
    },
  },
 
  computedFields: {
    slug: {
      type: "string",
      resolve: (doc) =>
        doc._raw.sourceFileName
          // hello-world.mdx => hello-world
          .replace(/\.mdx$/, ""),
    },
  },
}))

Once the JSONs are generated, you will notice a new field, slug, added to the post -

{
  "_id": "posts/hello-world.mdx",
  "title": "Hello World",
  "slug": "hello-world",
  "body": {
    "raw": "\nHello world! This is my first post.\n",
    "code": "var Component=(()=>{var sr=Object.create;..."
  }
  // [...]
}

We can do a lot more with the computedFields. We will explore this in more detail in future posts.

Let's tie it all together

Let's build a dynamic article page with the information available. The aim is to show the nicely formatted markdown content whenever the user navigates to the <base_url>/blog/hello-world page on the blog.

Create the following folder structure in the root of the project directory -

./
└── app/
    ├── blog/
    │   ├── [slug]/
    │   │   └── page.tsx
    │   └── page.tsx
    ├── layout.tsx
    └── page.tsx

We will use the page.tsx on line 5 to show a dynamic article page and page.tsx on line 6 to display a list of all blog posts.

Blog Posts List

Open up ./blog/page.tsx in your code editor and add the following code -

import Link from "next/link"
import { allPosts, Post } from "contentlayer/generated"
 
export default function ListPage() {
  // sort posts based on 'published' date
  const posts: Post[] = allPosts.sort((a, b) => {
    return compareDesc(new Date(a.published), new Date(b.published))
  })
 
  return (
    <div>
      <h1>Blog</h1>
      {posts.map((post) => (
        <div key={post.slug}>
          <h2>
            <Link href={`/blog/${post.slug}`}>{post.title}</Link>
          </h2>
        </div>
      ))}
    </div>
  )
}

Article Page

Once we click on any entry on the blog list page, it takes us to the /blog/<slug> page, where the content of the markdown file should be shown. Let's implement this functionality now -

import { allPosts, Post } from "contentlayer/generated"
import { getMDXComponent } from "next-contentlayer/hooks"
 
type Props = {
  params: { slug: string },
}
 
export default async function Page({ params }: Props) {
  const { slug } = params
  const post: Post | undefined = allPosts.find((post) => post.slug === slug)
  if (!post) {
    return <></>
  }
 
  const MdxContent = getMDXComponent(post.body.code)
 
  return (
    <div>
      <h1>{post.title}</h1>
      <section>
        <MdxContent />
      </section>
    </div>
  )
}

Contentlayer provides the getMDXComponent hook, which can render the mdx content on the page as HTML. Under the hood, our mdx file was transformed by Contentlayer using a bundler into JSX and cached in a JSON file, which was then picked up by NextJs to statically render using React.

Conclusion

Contentlayer greatly abstracts away the intricacies of using MDX content inside a NextJs project by doing most of the heavy lifting for us. It simplifies working with unified, remark and rehype ecosystems while providing interfaces to define a schema to our relatively freeform MDX content and type-safe data structures to access the transformed content.

It is also possible to pass custom MDX components to the MdxContent to enable the re-mapping of standard markdown constructs to React components. We will dive deeper into it in future posts of the series.

Pages that link to this

Note: Add reading time to blog articles

Share this article on Twitter