Ready-made Make.com automation - complete blogging system from topic research to publication

Today I'm sharing a ready-made Make.com automation with you. Instead of manually tracking news, writing articles, creating graphics, optimizing for SEO, and manually publishing on social media, I decided to delegate these tasks to machines. Below are three complementary scenarios that took over the entire blogging process – from finding topics to publishing on WordPress and automatic promotion across multiple platforms.

Scenario 1: Researcher – automatic trend aggregator

The first scenario starts every day and performs one task: finds current and potentially viral topics. In the described case, from the AI (artificial intelligence) industry.

Main modules and steps:

Perplexity AI module (model "sonar") – sends a query about the 3 most interesting topics from the previous day, from Polish and international social media, with emphasis on Rowan Cheung's profile on X.com.

RegExp Parser – extracts clean titles and topic descriptions from the response in the format |title – description|.

Aggregator and Feeder – converts results to array structure.

Google Sheets – Add Row – adds these topics to the spreadsheet as publication proposals.

The result? Every morning I have 3 fresh, inspiring topics from the last 24 hours in my Google spreadsheet – ready for review, acceptance, and then processing in the second scenario.

Scenario 2: Auto-editor – AI writes, edits, publishes

This scenario starts automatically when a new row appears in the spreadsheet. It's the central part of the process (pipeline) that creates a full-value blog post from a single sentence (!).

What does this scenario do?

1. Retrieves new topic from entries in Google Sheets prepared in the first scenario – google-sheets:watchRows module.

2. Orders Perplexity AI source search – we expand the given topic, checking GitHub, documentation, benchmarks, blogs, among others.

3. Passes results to OpenAI GPT-4 – prompt creates a ready SEO article in blog style, with Markdown headers, source links, section structure, neutral and concrete tone.

4. Generates post title – another OpenAI prompt creates a concise, SEO-friendly headline, e.g., "New AI features in Android 15".

5. Creates illustration prompt – text prompt in English generated for the FLUX model (Replicate).

6. Sends HTTP request to Replicate/FLUX 1.1 Pro – generates unique illustrative graphics.

7. Downloads graphic and creates media item in WordPress (wordpress:createMediaItem).

8. Publishes post in WordPress (wordpress:createPost) – automatically adds category, tag, author, "publish" status, content and graphics.

9. Generates filename from title – without Polish characters, up to 30 characters (e.g., aiwandroid15funkcje.webp).

10. Sends notification email about new publication – simple HTML message with link to the post.

This entire process takes just a few minutes and requires no intervention on my part. Each article is unique, based on current data, linguistically correct, and ready for promotion.

Scenario 3: Social – content distribution across all channels

The last scenario is the master of automatic promotion. After detecting a new WordPress publication (wordpress:watchPosts), it automatically creates and publishes posts on social media:

Channels and their handling:

Facebook Pages:

• OpenAI generates a post with a maximum of 5 emojis, first-person narrative and blog style.
• Post published with photo (CreatePostWithPhotos) and link as comment.

LinkedIn:

• An "article share" type post is created with thumbnail, headline and description (description based on post content).

Instagram:

• A post is sent with photo (graphic from blog) and caption same as on Facebook.

Reddit:

• Classic link post created to subreddit r/SubredditName.

Bluesky:

• A neutral, 220-character informational post is generated without emotional phrases.

Each platform receives a uniquely tailored version of the message — without repetition and content recycling. Everything is based on one blog post, but sounds different depending on the channel and its specifics.

Summary

With the help of Make.com, GPT-4, Perplexity AI, Replicate and WordPress, I created a hybrid, intelligent editorial system that can create and promote new blog posts daily — without human participation. Everything works fully automatically, based on current data and taking into account SEO and UX principles.

This is not just a proof-of-concept. Below I'm including ready-made blueprints for Make.com. You could say that thanks to these scenarios, you have a working editorial stack of the 21st century :)

An example of working automation based on these blueprints is the portal sztucznainteligencjablog.pl, where most posts are currently created automatically using the described scenarios.

Available blueprints:

• researcher.json - Trend research scenario
• autoredaktor.blueprint.json - Main content creation scenario
• sociale.blueprint.json - Automatic social media distribution