Blog post repurposed into four social posts with GPT

What it is

This n8n workflow turns one blog post into drafts for four networks. A webhook receives the post as JSON with a title, the text and an optional brand voice. A Code step checks that the text has at least 50 characters and gives the request an id. An OpenAI step reads the post and returns the key points, the core message, hashtags and SEO keywords. Four more OpenAI steps then write an X thread, a LinkedIn post, an Instagram caption and a Facebook post, one after another. A last Code step gathers them, and the result is added to a Google Sheet. The caller gets a short JSON reply. Eleven nodes run it, beside one sticky note.

Who it suits

Writers, marketers and small teams who publish long posts and want a first draft for each network. Skip it if you want posts published for you, because it only writes drafts into a sheet. Each run makes five OpenAI calls.

You need to know n8n’s settings: import a file, add two credentials, make a sheet and call a webhook.

Where to use it

Any n8n that can reach OpenAI and Google Sheets will run it, and the tool that sends the blog post must reach the webhook.

• n8n Cloud: import the file and publish it, then use the production webhook address
• Your own n8n server: the same file and steps, with the webhook reachable from the internet
• Google Sheets: one tab called Posts with seven column names in row 1
• OpenAI: the model that analyses the post and writes the drafts

Setup

Download saves a zip with workflow.json and Mindket’s setup notes in your downloads folder. Setup is two credentials and one sheet.

1. In n8n, create a workflow, open its menu, choose Import from File and pick workflow.json.
2. Add an OpenAI credential and select it in AI: Quick Analysis and in the four Generate steps.
3. In your Google Sheet, make a tab called Posts and type the seven headers in row 1.
4. Add a Google Sheets credential and select it in Save to Google Sheets. Replace <YOUR_SHEET_ID> with your sheet’s id.
5. Publish the workflow, then post a test blog post as JSON to the production address of Receive Content.

How to use it

Post each new article to the webhook. Read the four drafts in the sheet, edit them and publish them yourself.

Example

You post a blog post with a title, its text and a brand voice.

{"title": "Why we moved to self-hosting", "content": "The full text of the post, at least 50 characters", "brand_voice": "plain and friendly"}

The caller gets a JSON reply that says four drafts were made and to check the sheet. The sheet gets one new row with the X thread, the LinkedIn post, the Instagram caption and the Facebook post.

For newcomers and experts

If you are new. n8n draws a workflow as boxes joined by lines, and each box, a node, does one job. This one receives an article, asks an AI to find its main points and then asks it four times to write a post. Test with a short article first.

For experts. The four writers run in a chain, and each reads its facts from Merge Data.

Good to know

The webhook is open, and every call spends your OpenAI credit.

• Safety: anyone who finds the address can post to it. Switch the webhook’s authentication to a header check if the address leaves your own team
• Fixes: the original used an OpenAI node version that n8n does not have, so the five OpenAI steps are rewritten as the current OpenAI node on gpt-6-luna with JSON output. The temperature and the token limits are dropped, because small limits can cut a reasoning model’s answer short
• Order: the four writers now run one after another. In the original they all fed one Code step, which would have run once for each arriving branch
• Id: the original asked for a random id with a function that n8n’s Code node refuses by default. It now uses the time and a random string
• Sheet: the sheet id is a placeholder, the tab is called Posts, and the column list is spelled out so that n8n can write the row
• Quality: each writer asks for a length, such as under 280 characters for each tweet. Models can miss it, so check the drafts. Mindket did not check each network’s current length or hashtag rules
• Not run: Mindket did not run it against OpenAI or Google Sheets. The Code steps were tested only on sample data on a local computer
• Credentials: the JSON holds none