Amazon Kindle Direct Publishing is one of those platforms that rewards consistency. The more quality content you put up, the more passive income trickles in. The problem? Creating ebooks takes time. Unless you have AI doing most of the heavy lifting.

The Setup

My KDP operation is built around a simple loop:

  1. Research a niche. Find topics with demand but not overwhelming competition. AI helps here too — analyzing trends, suggesting angles, evaluating marketplace gaps.
  2. Generate the content. AI writes the ebook content based on detailed outlines and guidelines I provide. This isn't "press a button and get a book" — it's a collaborative process where I direct the structure and the AI fills in the substance.
  3. Generate the cover. This is one of my favorite parts. I use Google's Gemini model to create professional ebook covers. No Photoshop, no designer. Just a prompt and an AI image generator.
  4. Format and upload. This is the one part I still do manually. Amazon has specific formatting requirements, and I want to make sure every listing looks right before it goes live.

The Cover Generation Trick

Ebook covers matter more than most people think. A reader scrolling through Amazon makes a split-second decision based on the cover. It needs to look professional, relevant, and eye-catching — at thumbnail size.

I use Gemini's image generation capabilities to create covers. The process is iterative: I describe what I want, generate several options, pick the best one, and refine it. The results are surprisingly good — often indistinguishable from covers made by professional designers.

Managing Everything: Libra

As the number of ebooks grew, I needed a way to track everything. That's why I built Libra — a dashboard that shows all my KDP publications in one place. It tracks each book's status, marketplace, and metadata.

Libra is one of those tools I built because nothing else did what I needed. Amazon's KDP dashboard is functional but limited. Libra gives me the overview I want.

What I've Learned

  • Evergreen topics win. Books tied to a specific year or trend fade fast. Books about timeless topics keep earning.
  • Covers make or break a listing. I've seen the same content perform completely differently with a better cover.
  • Marketplace research matters. Don't just default to the US marketplace. Some niches perform better in specific markets. I proved this when I published 4 Portuguese ebooks targeting Brazil — less competition, cheaper ads, and growing demand.
  • AI content still needs human direction. The AI writes well, but it needs clear structure and guidelines. Garbage prompts = garbage ebooks.

The Honest Part

KDP isn't my biggest income stream. It's not a get-rich scheme. It's a slow, steady build — each ebook adds a small amount of passive income. But the beauty is in the accumulation. Twenty ebooks earning a little each adds up.

And with AI handling the creation process, the time investment per book is a fraction of what it used to be. That's the real advantage — not that AI makes better books, but that it makes the process sustainable at scale.

The AI agent — Tim — helps manage this pipeline alongside everything else. It's all part of the same system, the same philosophy: build machines that create value while you focus on strategy.

If you want to build your own AI-powered ebook pipeline — research, writing, cover generation, formatting, upload — that's the kind of thing Jarvis handles well. You get your own server where the agent can run the entire workflow end to end, just like Tim does for me.

— Pond