Automating LinkedIn AI Posts in 5 Minutes with n8n
An RSS trigger, an IF filter, and a GPT-4o-mini summary node chained together in n8n Cloud produce a fully automated LinkedIn post in about five minutes.
A LinkedIn post can be fully automated, from sourcing the news to writing the summary to actually publishing it, in about five minutes of setup, with no coding and no cost beyond a free trial. Here's the full workflow, built in n8n Cloud, step by step.
Why automate this at all
The idea started from browsing an n8n example gallery and noticing the range of things the tool could do, which led directly to trying a LinkedIn automation. The specific use case: there's a page followed daily for the latest AI updates, and rather than manually reading it, summarizing it, and posting about it, an LLM could read the content through its API, summarize it, and turn that summary directly into a LinkedIn post.
The workflow, piece by piece
The setup runs on n8n Cloud's free 14-day trial, which also comes with a batch of free OpenAI credits, roughly 100 credits in this case, enough to build and test the whole thing without spending anything.
RSS feed trigger. The workflow starts with an RSS Read trigger pulling from a specific feed URL, in this case an AI news feed. This trigger can run automatically whenever new content appears, and it's set to check on a recurring schedule, every 48 hours in this setup, so it always has the latest posts to work with.
Filtering for AI news. Next comes an IF node that checks the article's category field for whether it falls under an AI-related category. This step can hold more conditions if needed, but the setup here keeps it simple: one check, true or false. Testing the node shows exactly which branch, true or false, a given article falls into. Only articles matching the true branch move forward to the summarization step.
Summarizing with GPT-4o-mini. The next node is a chat completion call using OpenAI's free trial credits, specifically GPT-4o-mini, chosen because it's budget-friendly. The prompt requests a JSON output containing both a LinkedIn post and a URL. It asks for roughly a 100-word post, with line breaks between sections so the post formats cleanly once it's actually published, and instructions to keep the tone engaging and end with a question. The node uses a content snippet from the RSS item rather than the raw content field, because raw content often carries embedded images and HTML tags that would need to be stripped out first; the snippet avoids that mess entirely.
Posting to LinkedIn. The final node is LinkedIn's "create a post" action. Connecting a LinkedIn account is a one-time credential setup: click to create a new credential, connect the account, sign in, and authorize. The post is created as a person rather than an organization, since organization support wasn't needed for this use case. The post content field pulls directly from the LinkedIn post field generated by the summarization step, the media category is set to "article" since a link is being attached, and the original URL from the RSS item goes into the link field so the post directs readers back to the full article.
Testing and going live
Running the workflow as a test produces an actual LinkedIn post referencing the latest matching news item. From there, the prompt used in the summarization step is the main lever for tuning the output, adjusting tone, style, or the way the post opens can shape posts to match a specific voice rather than sounding generic. Switching from GPT-4o-mini to a larger model would also improve output quality, though the smaller model is enough for a working first pass. Once the workflow is activated, it runs on its own schedule, every 48 hours in this case, continuing to post automatically without further manual steps.
Key takeaways
- The full pipeline chains four n8n nodes: an RSS feed trigger, an IF filter for category matching, an OpenAI GPT-4o-mini summarization call, and a LinkedIn create-post action.
- n8n Cloud's free trial includes free OpenAI credits, enough to build and test the entire workflow at no cost.
- The summarization prompt requests JSON output with both a LinkedIn post and a source URL, targets about 100 words, and asks for line breaks and an ending question.
- Using a content snippet instead of raw RSS content avoids embedded HTML tags and images that would otherwise need separate cleanup.
- The workflow can be scheduled to run automatically, in this case every 48 hours, posting to LinkedIn without further manual intervention.
Try it yourself
If your organization wants a lightweight way to keep a LinkedIn presence active without manually posting every day, this same RSS-filter-summarize-post pattern works for any topic feed, not just AI news. This tutorial is part of Humanitarians AI's practical automation series, built for anyone with no prior n8n experience.
Chapters
Full transcript(auto-generated, with timestamps)
The automated post drops!
[0:04]Hi. So, the post, the LinkedIn post which you see in front of me was completely automated using a workflow and I'm going to show you exactly how I did it and it didn't take me more than 5 minutes to do this. So, I used N8 which is a secure workflow automation and as you can see in the example provided you could do a lot of fancy things. So I was just going through this GitHub and uh it just sparked why why don't I just try out an LinkedIn automation and I it was a pretty simple
Why LinkedIn automation rocks
[0:36]One but it's it's quite good. So what I did was so there's this page which I follow uh every day I just go through it uh to see about the latest updates. So I thought I would use uh their API. So, so this is the uh API link which I was using. So the idea was to read content from this and using any LLM uh summarize it and make it as a LinkedIn post. So that that was the basic idea. So I started
n8n Cloud workflow basics
[1:12]With uh logging into N810 and I set it up and workflow. This is the workflow. It's quite simple. when you could do a lot of other things as well in this but uh I just kept it quite simple and uh I didn't have any prior knowledge it didn't take a long time as well so this RSS free trigger is going to uh take details from this uh page so I set it like 2 days once so that it always has the latest and some recent post and I just use an if to check if the the news which I'm taking is an AI related news uh if it falls under an AI category and then I just use an open AI model to just summarize it. So I didn't pay for any of it here. So I'm just using the NA n free trial of 14 days and they did provide an open AAI free credits as well uh for uh I think they provided like 100 credits. So I'm just going to do the same setup and show you how you can do it as well. So I'll just start with creating a new setup. And the
RSS feed trigger setup
[2:22]First thing like I said let's add our RSS feed trigger. We can use our RSS read. So when we use our uh trigger what we can do is we can have like an automatic trigger whenever a new news is news is updated we could always trigger it. And in our RSS feed trigger. So this is the example feed URL which they have given. So I can see so but what I am using is this one. So right so this is getting the latest uh news and see you can see some categories over here. So this is what I use to filter out if it's part of an AI category. So I'm just going to keep it uh a custom or every 48 hours right so that every 48 hours this is going to run and going to do a post for me.
Filtering for AI news
[3:26]Okay so first RSS trigger is done and the next part is RF. So what I'm doing here is uh I'm just going to do some drag drop over here. Just take a categories and going to check if it contains AI. So I can have more conditions as well. So I'm just keeping it simple now. Uh when I do a test setup, I can see it's going to test this condition and give me like a true branch and a false branch. So only if it's true I'm going to go ahead with my open AAI models and uh carry on. So
GPT-4o-mini summary magic
[4:07]Next part is adding an open AI model. So what I'm going to do is just a chart completion. So I'm using the OpenAI free credits which enate and provides and it's just a text. So part of their free credits I think we are limited by using only few models. So I'll just go with GBT 4 mini which is very budget friendly. And about the prompt I have a prompt which I used uh in my previous post. So I'll show you how a word it says as well like a I want an JSON output which has LinkedIn post and URL I'll show you why I'm having getting this URL also and uh I want like a 100word LinkedIn post it should have like bras between them so that uh
You know it'll have like a better formatting when we do a LinkedIn post and we have it uh I'm just mentioning it to keep this engaging and end with a question. Okay. So, you can see all these things over here. It's quite simple. You can just uh drag drop this content snippet over here. And the reason why I'm not dragging the content is because sometimes the content has like uh you know uh all these uh attached images, this HTML tags uh like this. See, so it's always better when we have a content snippet uh like so I'll just give output content as a JSON and let's test the setup. Right? So this is the output which we're getting a LinkedIn post and a URL and the next step is
Posting to LinkedIn
[6:12]Linking it with our LinkedIn account. So they do have like a create a post action. So I've already uh linked in my account. Uh if you have to do yours, you can just click on create a new uh credential. Uh I don't think I need the organization support. So next thing you can you should just click on connect to your account, enter your password, sign in and then it's going to show you. So this is the previous one which I have linked and we're going to uh create a post post as a person. It shows up my uh name over here and this here we should add in the post content. Let
Me just click on JSON. So for the post content I'll just drag drop LinkedIn post. For the media category, I'm going to keep it article because we're going to attach a link in it as well so that the user can redirect to the uh article and complete reading the complete article on the topic. Right. I'm just going to add a feed for original URL. That's where our URL goes. Awesome. So yeah, so now we have a complete web flow and what I'm going to do now is do a test workflow. So it's going to just run and we are going to have a LinkedIn post. Just going to refresh my page. Right. So,
Live post reveal
[8:10]Here is a lengthen post and we do have a link and it's referencing your latest news. You can always like uh change your prompt and make it little better. Uh you know if you always post um little funny or you might have a good way how you interact you can just uh write those in your prompt and have uh better LinkedIn post tailored to you. And also I'm just using a GBT4 or mini model. you're going to have a far better outputs and results when you use GBT4 model. But it's it was just a small test. It didn't take more than five minutes for me to just put this all together. And this is my first time using it. Once if you're spending more time in it, you can like you know uh go around all the other tools which they has have to offer and you can build up like a awesome uh output. Right? So if I just activate the
Tips to tweak it
[9:23]Uh workflow and keep it. So every 48 hours it's going to get triggered and it's going to keep executing and you you're going to just be posting LinkedIn post automatically. Yeah. So that is all for now. Thank you.
More from INFO7375 Branding & AI
6:154 Scaling & Testing Your n8n Workflow with HAILogo
6:124 - Scaling & Testing Your n8n Workflow
2:31Professor Bear Explains: How to Use Claude AI to Create Any n8n Workflow
4:581 - Getting Started n8n
4:543 Datacleaning n8n with HAILogo
14:52