Content

Transform Blog Drafts into CMS-Ready Posts at Scale

Explore the use of Moonlit and ChatGPT in transforming rough drafts into polished blog content, ready for CMS publishing at scale.

Mohammad Abdin

Founder

February 9, 2024

This Article will explore a simple yet powerful use case for Moonlit. It's been more than a year since the release of ChatGPT and since then everyone has been exploring all the multitude of use cases that can be put to action using this powerful AI. One conclusion has been consistent, these language models are not optimal for long form content or more precisely not automation blog creation start to finish. However, they are really good at proof reading and improving existing content by being concise and following a clean structure as well as generating short-form content.

Today I will put these strengths into action by building a Moonlit App that takes a rough first draft for a blog post along with a short list of target keywords and have it proof read and optimized for a better reading experience as well as generate all the properties for our CMS. These can differ depending on your blog, for this example we will assume the CMS blog collection has the following fields:

  • Meta Title
  • Meta Description
  • Read Time (in minutes)
  • Tags

You can try it and clone it here

Building our App

I started by adding two text input nodes for getting the content copy and the keywords. Then our logic is quite simple we just used two ChatGPT nodes, the first one will be responsible for Proof Reading, it will take our rough draft and improve it using the following guidelines:

  • Maintain all the existing ideas in the draft
  • Rewrite sentences to be more concise and well-written.
  • Maintain the general tone and brand of the content
  • Optimize it for the following keywords without forcefully injecting them: {{text_input_627437}}
  • Return the content in markdown format.

The next GPT Node will be responsible for generating the CMS properties for our blog. So to make sure that we are getting a valid JSON response I'll tick the 'Force JSON' checkbox and also emphasize it in my prompt.

Making Custom Edits

I've set this app to allow cloning for anyone so you can clone it into your project and make your own custom changes (app). Every CMS uses a different set of properties so it only makes sense to have this app be shared publicly for cloning so you can copy it to your project to either:

  • Edit the prompts
  • Add/remove properties
  • Make the content fit your brand specifications
  • Change the output format

Or whatever change you need to fit your use-case!

Running it at Scale

This does save us a bit of time as it is, but what if we have tens of rough drafts or hundreds of blog posts that can be recycled? Then we can put the Bulk Run feature into action. I've prepared a short list for demonstration, a CSV containing 5 technical blog posts I wrote a few years ago.

In the bulk run tab I've created a new job and uploaded a CSV containing the keywords along with the blog content and started the job. I can now download the result then import into our CMS collection to get a list of blog ready for publishing immediately!

Written By

Mohammad Abdin

Mohammad is a full-stack developer, and the founder of Moonlit Platform. He holds a Bachelor's degree in Computer Science & Artificial Intelligence, and is committed to continuous learning and skill enhancement. His journey is marked by a steadfast dedication to developing and delivering exceptional product experiences.

4

Minute Read