DEV Community

tamay erdogdu
tamay erdogdu

Posted on

How I Started Making Passive Income With AI-Generated Shorts

How I Started Making Passive Income With AI-Generated Shorts

I’ve always been that person who spends evenings scrolling through TikTok and YouTube Shorts, wondering “How the heck do they crank out so many videos?” The answer? A mix of creativity, hustle, and lately… AI video automation. When I first heard about “AI Shorts” I was skeptical. Could a script‑writing bot really replace the brainstorming sessions I love? Could automated video production actually generate any revenue? This post is my honest journal of the past 45 days—wins, frustrations, and the tiny victories that now give me a modest stream of passive income AI.


The Problem: Too Little Time, Too Many Content Ideas

I work full‑time as a front‑end developer, and on weekends I dabble in side projects. My biggest pain point was content automation: I wanted to share quick coding tips, but filming, editing, and uploading each short took an hour. Even if I posted three times a week, the effort didn’t match the reach. I needed a system that could:

  1. Generate a concise script from a prompt.
  2. Pull relevant images or short clips.
  3. Add a natural‑sounding voiceover.
  4. Render a 15‑second video ready for YouTube Shorts, TikTok, and Instagram Reels.
  5. Post automatically without manual clicks.

If I could nail those five steps, I could finally treat content creation like a passive‑income machine.


Week 1: Diving Into n8n and the First Workflow

I started by reading about n8n workflow platforms. n8n is an open‑source automation tool that feels like a visual version of Zapier, but you can self‑host and add custom nodes. After a quick tutorial, I built a simple workflow that:

  • Took a prompt (e.g., “Explain CSS Grid in 30 seconds”).
  • Sent it to OpenAI’s GPT‑4 for a script.
  • Saved the script to a Google Sheet.

The first run was a disaster. The AI produced a 200‑word script—far too long for a short. I had to manually trim it, which defeated the purpose of automation. Mini setback #1: I spent almost three hours tweaking the prompt until the output hit the sweet 30‑word range.

Lesson learned: prompt engineering is a skill in itself. I added a “word‑count” node that re‑asks GPT if the script exceeds 45 words. After a couple of iterations, the workflow began to respect the length constraint.


Week 2: Adding Image Search and Voiceover

Next, I needed visual assets. I integrated the Unsplash API to fetch royalty‑free images based on keywords extracted from the script. The workflow now looked like this:

  1. Generate script.
  2. Extract keywords (using a simple regex node).
  3. Search Unsplash for matching images.
  4. Download the top result.

The image selection was hit‑or‑miss. Some keywords (like “flexbox”) returned abstract photos that didn’t convey the concept. Mini setback #2: I spent a day manually curating a fallback keyword list for technical terms.

For voiceover, I tried ElevenLabs but hit API limits in the free tier. I switched to a cheaper TTS service that still sounded natural enough for Shorts. Adding the TTS node was straightforward, but the audio length sometimes overshot the 15‑second window, causing the final video to be clipped. I added a “duration check” node that trims the script if the resulting audio exceeds 12 seconds.


Week 3: Automated Video Production and Scheduling

With script, image, and voiceover ready, the next piece was stitching them together. I used the FFmpeg node in n8n to overlay the image, sync the audio, and add a simple animated text subtitle. The command looked messy, but after a few test runs I got a clean 1080×1920 MP4 file ready for Shorts.

Now the final, satisfying part: auto‑posting. I connected the workflow to the YouTube Data API, TikTok’s upload endpoint (via a third‑party service), and Instagram Graph API. Each platform required its own authentication, but once set up, the workflow could publish the video instantly after generation.

By the end of week three, I had a fully automated pipeline that turned a single prompt into a published Shorts video across three platforms—exactly the content automation I was craving.


Week 4–5: Real Results and Passive Income AI

I decided to test the system with a modest schedule: three prompts per week (Monday, Wednesday, Friday). The topics ranged from “Quick React Hook tricks” to “CSS variables in 15 seconds.” After two weeks, the analytics looked promising:

  • YouTube Shorts: average 1.2k views per video, 5% watch‑through rate.
  • TikTok: 800–1k views, occasional spikes to 3k when the thumbnail caught attention.
  • Instagram Reels: 600–900 views, steady follower growth.

Monetization kicked in via the YouTube Shorts fund (around $0.30 per 1,000 views) and TikTok’s Creator Fund (roughly $0.02 per 1,000 views). Combined, the three videos generated about $2.50 per week—nothing huge, but a clear sign that passive income AI can be real.

The biggest surprise? The workflow ran entirely on my local machine overnight, consuming <5% CPU. I could set it and forget it, freeing my weekends for other projects.


What I Learned: The Good, the Bad, and the Realistic

  • Prompt engineering matters. A well‑crafted prompt reduces manual trimming.
  • Keyword extraction is key for relevant images; a fallback list saves time.
  • Audio length controls prevent clipping in the final video.
  • API limits can be a bottleneck; budgeting for a modest paid tier pays off.
  • Automation isn’t magic. You still need to monitor performance and tweak content for each platform’s audience.

Overall, the experience proved that AI video automation is more than a hype buzzword. It’s a practical tool that, when paired with an n8n workflow, can turn a tiny idea into a steady stream of content—and a modest side income.


My Recommendation

If you’re a developer or creator who wants to experiment with AI Shorts without spending weeks building a custom pipeline, I highly recommend trying a ready‑made solution. 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 spin, tweak the prompts to your niche, and watch the automation do the heavy lifting. Even if the income is modest at first, the time you save is priceless—and that’s the real win for any side‑project hustle. Happy automating!

Top comments (0)