Geddon Labs

Geddon Labs Geddon Labs specializes in training Stable Diffusion AI models to probe, document, and analyze alternate timelines and nonstandard realities.

Geddon Labs is a research division dedicated to the systematic exploration and documentation of alternate realities through the training of Stable Diffusion artificial intelligence models. Our primary objective is to develop, refine, and deploy AI models capable of interfacing with the informational static that defines the multiverse, with the intent to extract and visualize data from non-canonica

l realities. Standard operating procedure involves the curation of specialized datasets, iterative model training, and the subsequent deployment of these models to generate visual artifacts. These artifacts are catalogued and analyzed as potential evidence of cross-timeline interaction, anomalous phenomena, or previously undocumented universes. The general public is advised that all findings are preliminary and subject to further review. Geddon Labs remains committed to the responsible investigation of multiversal phenomena and the expansion of humanity’s understanding of the boundaries of reality.

Animatronics Generator v2.3 is live on CivitAI.Step into the Animatronic Universe. Brass joints and painted grins. Eyes ...
11/06/2025

Animatronics Generator v2.3 is live on CivitAI.

Step into the Animatronic Universe. Brass joints and painted grins. Eyes that track from darkened stages. The crackle of servos, the hum of circuitry coming back to life. Fur worn smooth by ten thousand hands. Metal creased by decades of motion.

Download the model. Generate new creatures. Bring something back from the arcade that shouldn't exist—but does, because you created it.

The threshold is now open.

https://civitai.com/models/1408208/animatronics-style-or-flux1d

After hours of recursive tuning in the heart of Geddon Labs, our team is excited to announce the public release of the n...
11/04/2025

After hours of recursive tuning in the heart of Geddon Labs, our team is excited to announce the public release of the new Dambo Troll Generator v2—now available for download exclusively on Civitai.

This update is more than a version bump. Higher-resolution imaging arrays and refined semantic containment procedures have altered the way Dambo trolls emerge in your render field. Artifacts like wood, soap, concrete, and plastic carry specific, testable signatures of its material and imagined construction. Bark splinters stand apart from plastic seams; light refracts through translucent soap arms when specified, pooling on virtual counterspace.

Most importantly, the new captioning protocol encodes place, purpose, and presence. Write “plastic Dambo troll, assembled from translucent bottles, standing in a sunlit playground,” and the model deploys a definitive locus—object, material, atmosphere, all in precise alignment. Concrete trolls reliably surface at hazard-taped construction sites; paper trolls wander bustling mills, not arbitrary landscapes.

Ready to explore? Download Dambo Troll Generatorv2 now on Civitai and begin your own survey of the ever-growing Dambo Dimension.

https://civitai.com/models/1818617?modelVersionId=2376348

Announcing the Kaijin Generator LoRA v2.3 for Qwen Image! The updated version was trained exclusively on 200 image of Ja...
11/03/2025

Announcing the Kaijin Generator LoRA v2.3 for Qwen Image!

The updated version was trained exclusively on 200 image of Japanese kaijin—there were zero Batman, Spiderman, or other American superheroes anywhere in the data. Yet, when prompted with archetypal traits and thematic cues, the model generated images unmistakeably resembling icons like Spiderman, Batman, Hawkgirl, and Rorschach—including distinctive details and visual signatures—all from tokusatsu-inspired prompts.​

This is not random chance or mere visual remixing. These results strongly suggest that the superhero pattern is a universal attractor in the AI’s latent geometry—a reality archetype capable of emerging from underlying symbolic cues and structure, independent of specific training images.

Geddon Labs invites you to experience this revelation firsthand. Download the Kaijin Generator LoRA v2.3 for Qwen Image (available now on CivitAI), experiment with your own prompts, and see which superhero archetypes manifest for you.​

https://civitai.com/models/2047514

Ever have one of those weeks where a mysterious new creature takes over your creative universe? Until recently, we’d nev...
10/29/2025

Ever have one of those weeks where a mysterious new creature takes over your creative universe? Until recently, we’d never even heard of Labubu dolls. But then a kindly gentleman named Bingo from the podcast halfwayokay opened our eyes, and suddenly, these wild plush monsters are everywhere!

Swipe through to see our favorite samples from the Labubu Generator. Each doll with their own surprise and story, multiplying like tribbles in your imagination. We started with a handful... now we’re surrounded by a parade of mischievous Labubus, and honestly, we wouldn’t have it any other way.

Curious? Want your own Labubu invasion? Visit our Civitai page and download the Labubu Generator model. Give it a prompt, turn up the strength, and let the Labubu fever take over—just like it did for us!

https://civitai.com/models/2082902/labubu-generator-or-qwen-image-lora

The next evolution in diffusion models is here! Our upgraded Animatronic Generator powered by Qwen Image offers a portal...
10/29/2025

The next evolution in diffusion models is here! Our upgraded Animatronic Generator powered by Qwen Image offers a portal into the extraordinary.

With this release, you can generate animatronic creatures and worlds that feel uncanny and vivid, revealing new layers of the animatronic universe. The final model is now available for download on Civitai. See what emerges when you peer inside.

Download today and help shape this new dimension: https://civitai.com/models/1408208/animatronics-style-or-flux1d

We’re in the final stretch on our new Animatronic Generator model. There are a few technical refinements left, but the e...
10/27/2025

We’re in the final stretch on our new Animatronic Generator model. There are a few technical refinements left, but the early results are promising—fresh animatronic scenes emerging from the threshold, each a glimpse of what’s soon to manifest.

Here’s a sample of what’s possible. Full release is close; thanks for watching as we tune reality between potential and form.

Geddon Labs is back! After a short dive into the unknown, we’ve finally cracked the code: our new Qwen Image LoRA models...
10/27/2025

Geddon Labs is back! After a short dive into the unknown, we’ve finally cracked the code: our new Qwen Image LoRA models can see farther and sharper into other worlds than ever before.

Here's a few images of our lead researcher exploring strange new worlds, mystical forests, spectral deserts, and hidden dreamlands. Each snapshot is a message from another reality, a glimpse into realms beyond the horizon.

This is just the beginning. Where will our models take us next?

Take a trip through liminal space in the new Geddon Labs RPG. Available exclusively on Infinite Worlds! https://infinite...
08/07/2025

Take a trip through liminal space in the new Geddon Labs RPG. Available exclusively on Infinite Worlds!

https://infiniteworlds.app/

An AI-based adventure game

Multiversal signatures of large wooden humanoids, congruent with works attributed to artist Thomas Dambo, have manifeste...
07/29/2025

Multiversal signatures of large wooden humanoids, congruent with works attributed to artist Thomas Dambo, have manifested in various global loci. With confirmation of cross-timeline constructs, Geddon Labs has commissioned the Dambo Troll Generator model—an experiment in reality extraction and visual artifact generation.

Address

Springboro, OH

Alerts

Be the first to know and let us send you an email when Geddon Labs posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Shortcuts

Share

Category