A New Chapter: Why I’m Leaving CivitAI Behind and What Comes Next

January 2026

If you’ve followed my work over the past months, you know that creating ComfyUI workflows isn’t just something I do—it’s something I genuinely love. There’s a unique satisfaction in taking a complex chain of nodes, fine-tuning every parameter, and watching it produce images that blur the line between artificial and real. That passion hasn’t changed. But my approach to sharing that work is evolving.

With the arrival of 2026, I’ve made a decision that’s been brewing for a while: I’m stepping away from CivitAI and similar platforms. No more updates there. No more splitting my attention across a dozen different sites. Instead, I’m consolidating everything here—on my website and my GitHub repository.

Why Now?

Late last year, I closed my Patreon page. It was a difficult choice, but ultimately the right one. Managing multiple platforms, each with its own expectations, formats, and communities, had become more about maintenance than creation. I found myself spending hours reformatting the same workflow for different sites, writing variations of the same descriptions, answering the same questions in different comment sections.

That’s not why I got into this.

I got into workflow development because I wanted to push boundaries. I wanted to explore what’s possible when you chain together the right models, the right upscalers, the right enhancement techniques. I wanted to build pipelines that don’t just work, but work beautifully—reliably producing results that rival professional photography.

So I’m simplifying. One website. One GitHub. One place where everything lives, properly documented, consistently updated.

The Models That Matter

Here’s the thing about the AI image generation landscape: it moves fast. Absurdly fast. New models drop every week. New techniques emerge. Yesterday’s state-of-the-art becomes today’s baseline.

But not everything that’s new is worth chasing. After months of testing, breaking, rebuilding, and refining workflows, I’ve identified a core set of models that I believe represent the current peak for photorealistic generation. These are the foundations I’ll be building on:

Z-Image Turbo is my new foundation for photorealistic generation. Developed by Tongyi-MAI, this model delivers exceptional quality with remarkable speed. Its ability to render natural skin tones, realistic lighting, and coherent compositions makes it ideal for production workflows where both quality and efficiency matter. The “Turbo” architecture means faster iterations without sacrificing the detail that photorealism demands.

Illustrious has earned a permanent place in my pipeline for its remarkable handling of fine details—skin texture, fabric weave, the subtle interplay of light and shadow. When you’re aiming for hyperrealism, these details are everything. It excels particularly in portrait work, where the difference between “good” and “believable” often comes down to microscopic elements that other models overlook.

WAN 2.2 has become my go-to for video generation. The temporal consistency, the smooth motion, the way it handles transitions—it’s the best tool I’ve found for bringing static concepts into motion. As AI-generated content increasingly moves toward video, having a reliable motion pipeline isn’t optional anymore. It’s essential.

Three models. That’s it. Not because others aren’t capable, but because focus beats fragmentation. I’d rather master three tools completely than dabble in a dozen.

What You’ll Find on GitHub

My GitHub repository is becoming the central hub for everything I create. Here’s what that means in practice:

Complete, Production-Ready Workflows. Not experiments. Not half-finished concepts. Every workflow I publish will be tested, documented, and ready to use. You download it, you load it into ComfyUI, it works.

Detailed Documentation. Every workflow comes with a comprehensive README: required models, custom nodes, installation steps, recommended settings, troubleshooting guides. No more guessing why something isn’t working.

Version Control. Git wasn’t designed for workflow development, but it’s remarkably effective. You’ll be able to see exactly what changed between versions, roll back if something breaks, and understand the evolution of each pipeline.

Transparency. When I optimize a workflow, you’ll see the commit history. When I fix a bug, it’s documented. When I add a feature, the reasoning is there in the notes. This isn’t a black box—it’s open development.

The Philosophy Behind the Change

There’s a broader principle at work here, and I want to be transparent about it.

The AI generation community has grown explosively, and that growth has brought challenges. Platforms like CivitAI serve an important purpose—they lower the barrier to entry, they create discoverability, they build community. But they also create fragmentation. They encourage quantity over quality. They make it harder to maintain coherent, well-documented projects.

I want to do fewer things, but do them exceptionally well.

I want every workflow I release to be something I’m genuinely proud of. Something that represents the best of what’s currently possible. Something that comes with the documentation and support that serious creators need.

That’s harder to do when you’re scattered across five platforms. It’s much easier when everything lives in one place, under your direct control.

How to Stay Connected

If you want to follow my work going forward, here’s what you need to know:

This website—tenofas.ai—is home base. Blog posts, announcements, tutorials, and guides will all live here. If you’re reading this, you’re already in the right place.

GitHub (github.com/Tenofas) is where the workflows live. Star the repositories you’re interested in, and you’ll get notifications when updates drop. Every major release will also be announced here on the blog.

X/Twitter (@tenofaz) is where I share quick updates, work-in-progress previews, and engage with the community. It’s the fastest way to see what I’m working on and to ask questions.

That’s it. Three places. No Patreon tiers to navigate. No platform-specific formatting to decode. No wondering which version of a workflow is the most current.

A Personal Note

I started this journey because I was fascinated by what AI could create. I continue it because I’ve found a community of people who share that fascination—people who want to understand not just what works, but why it works.

Closing Patreon was hard. Stepping away from CivitAI feels strange after so many months of activity there. But I genuinely believe this consolidation will result in better work, better documentation, and ultimately more value for the people who use my workflows.

Thank you for being here. Thank you for your support, your feedback, your patience when things don’t work perfectly. Thank you for pushing me to be better.

2026 is going to be interesting. Let’s see what we can build.

—Tenofas