I don’t decide what content to post anymore. I let a workflow do it for me.
Every morning, I wake up to an email. It tells me what’s trending. It gives me content ideas based on real data. And it logs everything into Airtable so I never lose an idea.
No guessing. No doom-scrolling for inspiration. Just data.
Why I Built This Workflow
I’m Charles Dove. I run CC Strategic, an AI automation agency, and I build in public over at Charlie Automates on YouTube.
Here’s the problem I was facing. I’d sit down to create content and spend 30 minutes just figuring out what to talk about. Should I cover this tool? Should I react to that news? Is anyone even searching for this topic right now?
That’s wasted time. And as someone running an agency, wasted time is wasted money.
So I built an n8n workflow that scrapes YouTube, X (Twitter), and Perplexity every single day. It pulls trending data from each platform, runs it through AI, and delivers two things: a daily email report and a fresh batch of content ideas in Airtable.
The workflow does three things at once. It pulls top videos from YouTube about AI automation, grabs the latest tweets from key accounts, and runs a Perplexity search for trending news. Then it merges everything, generates a report, brainstorms content ideas, and sends it all to me before I’ve had my first cup of coffee.
The Three Data Sources
YouTube Scraping
The first part of the workflow connects to YouTube. It searches for 10 videos about AI automation and pulls the data.
Here’s where it gets interesting. I use an Apify actor to grab the transcripts of those videos. If you’re not familiar with Apify, it’s a platform full of pre-built scrapers. Google Maps scraping, website scraping, you name it. In this case, I’m using a specific actor just for YouTube transcripts.
Once the transcripts come in, the workflow runs them through an AI model. I used GPT-4.1 mini for this part because it’s cheap and fast. The AI summarizes each video and gives me a title, length, quick summary, and deep dive.
All of that data gets pushed into my Airtable base. I have a custom tab just for YouTube video summaries. So every day, I can glance at what’s trending and see condensed breakdowns of what other creators are talking about.
The workflow also filters out YouTube Shorts based on video duration. I only want long-form content because that’s where the real substance lives.
X (Twitter) Scraping
The second branch of the workflow pulls tweets. Same approach as YouTube, but I’m using a different Apify actor called Tweet Scraper v2.
It pulls the latest tweets, aggregates all the data, and runs them through the same AI pipeline. The AI model summarizes the tweets, parses the key information, and splits each tweet into its own entry.
Then all of that goes into Airtable, just like the YouTube data. Different tab, same base. Everything in one place.
One thing I want to call out. For the AI models that summarize the YouTube transcripts and tweets, I only used GPT-4.1 mini. It’s cheap. It gets the job done for simple summarization. You don’t need a top-tier model for parsing and organizing data.
Perplexity Search
The third branch is the Perplexity search. I have it set to ask: “What are the top three trending news stories in AI and AI automation today?”
That question changes based on what you’re tracking. If you’re in real estate, you’d ask about real estate trends. If you’re in SaaS, ask about SaaS. The point is it gives you a pulse on your industry pulled from the live web.
And here’s the cost breakdown. Apify is maybe $35 a month. That covers all the credits for each actor’s run. Sometimes it costs three cents per run. Sometimes less. Perplexity is also very inexpensive. You’re not breaking the bank to run this daily.
Merging Data and Generating the Report
This is where it all comes together.
After the three branches finish pulling data, the workflow merges everything. YouTube summaries. Tweet summaries. Perplexity news. Three outputs, all connecting and appending into one aggregated dataset.
Then it sends that combined data through two separate AI processes.
The Report Generator
The first process is the report generator. This is what creates the daily email I wake up to every morning.
The AI takes all the merged data and generates a report in HTML format so the email looks clean. It actually looks like a real newsletter was sent to me.
My daily report includes:
- TL;DR section with a high-level breakdown of what the report covers
- YouTube trends showing what’s trending in my niche
- X (Twitter) trends showing what people are talking about
- Perplexity news with the latest AI and automation headlines
- Key content ideas for the week pulled from all three sources
For this part, I used a higher-tier model. I ran GPT-5 through OpenRouter because the report is the thing I actually read every day. I want it to be sharp and insightful.
The Content Brainstormer
The second process is the content brainstorm node. This is where the real magic happens.
The AI takes all the aggregated data and generates specific content ideas. The prompt tells it how to act as my content strategist, what kind of ideas I’m looking for, and how to format the output.
The key rule: output valid JSON. That way the data flows cleanly into the next nodes without breaking. The ideas get parsed, split into individual entries, and then pushed into my Airtable content ideas board.
So every single day, I have fresh ideas waiting for me. Generated from real data. Not from my own head at 6 AM.
How I Actually Use the Ideas
Here’s the part most people miss. The workflow doesn’t just hand you ideas and walk away. You still need to decide what to create. But the decision is a thousand times easier when you have data backing it up.
When I’m ready to create content, whether it’s long-form or short-form, here’s what I do. I’ll open my Airtable, grab the newest content ideas, copy them, and open a new chat in my ChatGPT project.
Then I’ll paste the ideas and say something like: “Based on these ideas, give me five short-form ideas and one long-form idea.”
It runs. It delivers. And now I have a clear plan for the day.
If I want to go deeper on any idea, I just say “let’s expound on this one.” The AI gives me the structure, a simple script if I choose to use it. Then I can create the content however I want. I can use an AI avatar, do a raw video like the one I recorded for this tutorial, or write a blog post.
You could take it a step further and create a custom GPT trained to view this data from the workflow every single day. You could probably even automate that part too.
The point is this: you’re not guessing anymore. This is pure data hunting.
The Full Workflow Breakdown
Here’s the step-by-step of how the entire n8n workflow flows:
- Trigger fires on a daily schedule
- YouTube branch searches for videos, pulls transcripts via Apify, runs AI summary, logs to Airtable
- X (Twitter) branch pulls tweets via Apify, runs AI summary, logs to Airtable
- Perplexity branch searches for trending news, formats the output as JSON
- Merge node combines all three data outputs
- Aggregate node consolidates into a single dataset
- Report generator creates HTML email report using GPT-5 via OpenRouter
- Email node sends the report to your inbox
- Content brainstormer generates content ideas using GPT-5 via OpenRouter
- Parse and split breaks ideas into individual entries
- Airtable node logs each idea into your content ideas board
Tools You Need
| Tool | Purpose | Cost |
|---|---|---|
| n8n | Workflow automation engine | Free (self-hosted) or paid cloud |
| Apify | Web scraping (YouTube transcripts, tweets) | ~$35/month |
| Perplexity API | Live web search for trending news | Low cost per query |
| OpenRouter | Access to GPT-5 and other models | Pay per token |
| Airtable | Database for storing ideas and summaries | Free tier available |
| Gmail/SMTP | Sending the daily report email | Free |
Why This Beats Manual Research
Let me be real. Before this workflow, I was spending 30 to 45 minutes every morning scrolling YouTube, checking X, and reading articles. Just to figure out what to post.
Now I spend zero minutes on research. The workflow handles it overnight. When I wake up, the report is in my inbox and the ideas are in Airtable. I just pick one and start creating.
Over a month, that’s roughly 15 hours saved. Over a year? That’s 180+ hours. Almost 23 full work days. All reclaimed.
And here’s the thing. The ideas are better too. They’re based on what’s actually trending right now. Not what I think might be interesting. Real data beats gut feeling every time.
How to Get This Template
If you want to steal this exact workflow, join CC Strategic AI on Skool. You’ll get the n8n template, the Airtable base template, and access to the community where you can ask questions.
You’ll also get access to Saturday calls where I go deeper on builds like this. If you have questions about setting up the Apify API, connecting the different pieces, or customizing the prompts for your niche, that’s the place to get answers.
Customizing It for Your Niche
One thing I want to be clear about. What works for me won’t always work for you. My workflow is tuned for AI and automation content. But the structure is universal.
If you’re a fitness creator, swap the YouTube search terms to fitness topics. Change the Perplexity query to trending fitness news. Adjust the AI prompts to brainstorm fitness content ideas.
If you’re a real estate agent, search for real estate market updates. Pull tweets from real estate influencers. Ask Perplexity about housing market trends.
The bones of this workflow apply to any niche. You just change the inputs.
And the prompts? Customize them. What’s best for my business is not always going to be best for your clients and your viewers. Make it yours.
FAQ
How much does this workflow cost to run daily?
The total daily cost is minimal. Apify runs cost about three cents each. Perplexity queries are fractions of a cent. The biggest cost is the GPT-5 calls through OpenRouter, which might run a few cents per report. All in, you’re looking at maybe 10 to 20 cents per day. That’s roughly $6 a month for daily automated content research.
Do I need to self-host n8n to use this workflow?
No. You can use n8n Cloud if you don’t want to deal with Docker. But I personally run n8n locally in Docker because it gives me more control and zero recurring cost for the n8n platform itself. Setting up Docker takes 5 to 10 minutes.
Can I use a different AI model instead of GPT-5?
Absolutely. For the summarization steps, I use GPT-4.1 mini because it’s cheap and fast. For the report and brainstorm nodes, you can use Claude, Gemini, or any model available through OpenRouter. The workflow is model-agnostic. Just swap the API connection.
What if I’m not in the AI niche?
This workflow works for any niche. Just change three things: the YouTube search terms, the Perplexity search query, and the X (Twitter) accounts you’re scraping. The structure stays the same. Only the inputs change.
How long does the workflow take to run?
About 2 to 5 minutes depending on how many videos and tweets it’s pulling. The Apify actors need time to scrape, and then the AI models need a few seconds to process. But since it runs on a schedule, you never notice the wait. It runs while you sleep.
Do I need to know how to code?
No. The workflow uses n8n’s visual builder. The only “code” is a small filter to remove YouTube Shorts based on video duration. And that’s already built into the template. If you want to customize the AI prompts, that’s just plain English.
Can I add more data sources beyond YouTube, X, and Perplexity?
Yes. You could add Reddit scraping, LinkedIn post monitoring, Google Trends, or even RSS feeds from industry blogs. Each new source just becomes another branch that feeds into the merge node. The more signals you feed in, the better your content ideas get.
If you want to see this workflow in action, watch the full walkthrough on the Charlie Automates YouTube channel.
Ready to stop guessing and start using data? Join CC Strategic AI on Skool to grab the template.
Want hands-on help building this for your business? Work with me 1-on-1 or book a call with CC Strategic.