Back to DevLog

ImageStack Course Launch at $97

3 min read

ImageStack just went live at $97. It's a 6-module course that takes you from "I've heard of Stable Diffusion" to running production image generation workflows on RunPod with custom-trained LoRA models.

Why I built this

I've been generating AI images for client projects and my own products for over a year now. The gap between what's possible and what people actually know how to do is enormous. Most tutorials stop at "install ComfyUI and click generate." They don't cover:

  • How to train LoRA models on your own data
  • How to set up RunPod for cost-effective GPU access
  • How to build repeatable workflows that produce consistent results
  • How to go from a single image to a production pipeline

That's exactly what ImageStack covers.

The 6 modules

  1. Foundations — Understanding diffusion models, model types (SD 1.5, SDXL, Flux), and when to use each
  2. ComfyUI Deep Dive — Node-based workflows, custom nodes, and building reusable templates
  3. RunPod Setup — GPU instance selection, persistent storage, and cost optimization (I run most jobs on 3090s at $0.22/hr)
  4. LoRA Training — Dataset preparation, Kohya training scripts, and hyperparameter tuning
  5. Production Workflows — Batch generation, upscaling pipelines, and quality control
  6. Business Applications — Product photography, marketing assets, and client delivery workflows

The vibe-coding angle

I built the entire course platform using Claude Code in about 3 days. The landing page, the payment integration (Stripe), the content delivery — all vibe-coded. The course content itself took longer — about two weeks of recording, editing, and writing supplementary materials.

The irony of using AI to build a course about AI is not lost on me. But that's the whole point of the vibe-coding methodology: use the tools available to ship fast.

Pricing strategy

$97 is the launch price. It's deliberately accessible — I'd rather have 500 students at $97 than 50 at $497. The course pays for itself if you save even a few hours of GPU time from the optimization techniques in Module 3.

I'm not doing a free tier or a freemium model. The free content lives on the blog and YouTube. ImageStack is for people who want the structured, complete path.

Early reception

Soft-launched to my email list first. 23 sales in the first 48 hours, which is solid for a list of about 800 people. The feedback so far is that Module 4 (LoRA Training) is the most valuable — that's the part most people can't figure out on their own.

If you're interested in AI image generation and want a no-BS path from beginner to production, check it out. The link's on the projects page.

Share this post