v2.0 Beta
June 2026We rebuilt everything.
This isn't a patch. It's not a feature drop. We tore the entire platform down to the studs and rebuilt it from scratch — new engine, new models, new design, new everything. UD started as a scrappy experiment in uncensored AI art. Thousands of you showed up and made it something real. But the old platform was held together with duct tape and good intentions. You deserved better. So we rebuilt it. From the ground up. For you.
100+ models. Leading labs. Open source. All of it.
Before, we had 4 models. The new UD launches with over 100 — from the biggest names in AI research and the open source community alike. We're kicking things off with models like ByteDance's Seedream, OpenAI's GPT-Image, and Recraft, with more arriving every week. The jump in quality from what you remember is massive.
Influences. 20,000+ ways to shape your art.
This is brand new. Influences (powered by LoRAs) let you steer a model's output with styles, characters, concepts, and aesthetics created by the community. Think of them as creative lenses — stack them, blend them, make the model yours. We're launching with over 20,000 influences, and expect that number to grow rapidly. Browse them, preview them, apply them in one click.
Image-to-image. Start from what you have.
Another first for UD. Upload a reference image and use it as the starting point for your generation. Transform photos into illustrations, iterate on compositions, or use rough sketches as the seed for polished output. Pair it with influences for even more control over the final result.
Unlimited usage. New engine. No more tokens.
Everything under the hood is new. The old token system is gone. In its place: unlimited usage for most models — generate as much as you want, never blocked. A handful of premium models — to start, ByteDance's Seedream, OpenAI's GPT-Image, and Recraft's models — draw from a separate Heavy pool with its own generous weekly allowance. The UI is rebuilt from scratch — mobile-first, new model explorer, new generation workflow. Every pixel, reconsidered.
This is just the start.
We're calling this a beta because we're not done. The foundation is solid, but we're shipping fast. Expect new features, models, and improvements to roll out regularly.
Changes in v2.0 Beta
100+ AI models from leading labs and open source community
Kicking off with ByteDance Seedream, OpenAI GPT-Image, and Recraft. More models shipping weekly.
Influences — 20,000+ community LoRAs
Brand new feature. Browse, preview, and apply style influences to any model.
Image-to-image generation
Use reference images as starting points for generation. Combine with influences.
Unlimited usage for most models
Tokens are gone. Everyday models are never blocked; premium models draw from a separate Heavy pool with a weekly allowance.
Complete platform redesign
New UI, navigation, model explorer, and generation workflow. Mobile-first.
New model explorer
Browse, search, and discover models with detailed info and sample outputs.
Professional-tier image quality
Major quality upgrade across all models compared to previous generation.
ByteDance Seedream
High-quality image generation from ByteDance's latest research.
OpenAI GPT-Image
Premium image generation and editing from OpenAI.
Recraft
Premium image models from Recraft.