DEV Community

tamay erdogdu
tamay erdogdu

Posted on

The Workflow That Posts to 4 Platforms While I Sleep

#ai #automation #python #tutorial

The Workflow That Posts to 4 Platforms While I Sleep

I’ve been juggling a full‑time dev job, a side‑project vlog, and a tiny “passive income AI” experiment for the past year. Most evenings end with me staring at a blinking cursor, wondering how I could possibly squeeze another video into the day. The answer? A little bit of AI video automation and a lot of patience (and a lot of coffee).

Below is my week‑by‑week diary of building a n8n workflow that creates, renders, and publishes a short video to YouTube, TikTok, Instagram Reels, and even a niche community forum—while I’m asleep.

Why I Needed AI Video Automation

The problem: content fatigue

Every time I tried to add a new piece of content, I hit the same wall:

  1. Idea generation – I could think of topics, but turning them into a script was a slog.
  2. Voiceover – Recording my own voice sounded okay, but the quality varied.
  3. Video assembly – Adding images, syncing them to the narration, and exporting took hours.

I was skeptical that any tool could handle all these steps without a massive learning curve. Yet, the promise of automated video production kept nudging at me.

The goal: a “set it and forget it” pipeline

I wanted a single button that would:

  • Pull a trending keyword from Google Trends.
  • Generate a concise script using GPT‑4.
  • Find royalty‑free images that match each line.
  • Convert the script to a natural‑sound voiceover.
  • Stitch everything together into a 60‑second AI Shorts video.
  • Upload the final file to four platforms automatically.

If I could get this to run on a schedule, I could finally start creating “passive income AI” content without staying up until 3 a.m.

Building the n8n Workflow

Week 1 – Mapping the steps

I started by sketching a flowchart on a napkin. The core nodes turned out to be:

  1. Trigger – Cron node that runs every 24 hours.
  2. Keyword fetch – HTTP request to Google Trends API.
  3. Script generation – OpenAI node with a custom prompt.
  4. Image search – Unsplash API node with keyword filters.
  5. Voiceover – ElevenLabs node for realistic TTS.
  6. Video rendering – FFmpeg node that merges images, voice, and background music.
  7. Platform upload – Separate HTTP nodes for YouTube, TikTok, Instagram, and a Reddit‑style forum.

I was surprised at how clean the visual editor in n8n was. Drag‑and‑drop, set the credentials once, and you’re good to go.

Week 2 – First setbacks

The first setup took me longer than expected. Two issues stuck me in place:

  • Unsplash rate limits – The free tier only allowed 50 requests per hour, and my workflow was hitting it every run. I added a simple cache node to store the last image set, reducing calls by ~80 %.
  • YouTube OAuth – Getting the refresh token to auto‑renew was a nightmare. After digging through the docs, I switched to a service‑account approach that finally worked.

Even with these hiccups, the first successful run produced a 45‑second video that I manually uploaded to YouTube. The view count was modest, but the time saved was huge.

Week 3 – Polishing the content automation

Now that the skeleton was functional, I fine‑tuned the AI Shorts quality:

  • Prompt engineering – I added a “hook” line at the start of every script (“Did you know…?”) and a call‑to‑action at the end.
  • Image selection – I introduced a “color‑balance” filter to ensure the visuals matched the brand palette.
  • Voice tones – Switching from a default “neutral” voice to a “friendly” tone increased watch‑through rates by ~12 %.

I also added a small “error‑catch” node that emails me if any step fails, so I’m not left wondering why my post didn’t go live.

First Results: 30 Days of Automated Posting

After 30 days of running the workflow overnight, here’s what I observed:

Platform Avg. Views per Video Avg. Engagement Time Saved (hrs/week)
YouTube Shorts 1.2 k 4 % watch‑time 6
TikTok 2.5 k 6 % likes 5
Instagram Reels 1.8 k 5 % comments 4
Reddit‑style Forum 300 2 % upvotes 2

The numbers aren’t viral, but they’re consistent. More importantly, I’m now spending ≈12 hours/week on actual creative brainstorming instead of repetitive editing. The workflow is literally posting while I’m in bed, which feels like the definition of passive income AI for me.

A minor setback: platform policy changes

Mid‑month, TikTok updated its API, and my upload node started throwing “403 Forbidden” errors. A quick look at the changelog and a new authentication token fixed it in a few hours. This reminded me that any content automation system needs regular maintenance—nothing is set‑and‑forget forever.

What I Learned

  • Start small. I began with just YouTube; adding the other platforms later kept the scope manageable.
  • Embrace the bugs. Each error taught me something about API limits, credential handling, and the importance of logs.
  • Measure, don’t guess. The simple spreadsheet I kept for views and engagement helped me see the real ROI of the automation.

If you’re a developer who’s been curious about AI video automation but feels overwhelmed, I’d say: give n8n a spin. The visual workflow builder makes it far less intimidating than writing a custom Python script from scratch.

My Recommendation

For anyone looking to replicate this setup without spending weeks on trial‑and‑error, I recommend trying a pre‑built solution that already stitches together all the moving parts. The tool I’m using is called AI Shorts Factory (https://8622430312019.gumroad.com/l/gujqfy) — it’s an n8n workflow that costs $20 one‑time and handles everything: AI script generation, image search, voiceover, video production, and auto‑posting to YouTube, TikTok, and Instagram.

Give it a try, tweak the prompts to suit your niche, and you’ll have a reliable automated video production pipeline that works while you sleep. Happy automating!

Top comments (0)