Another small roundup of some of the more Lovecraftastic recent ‘style plugins’ for Stable Diffusion 1.5 AI image-generation. Free, and for use free local desktop software such as InvokeAI 5.x. You will however need a decent gaming graphics-card in your PC, ideally at least an NVIDIA 3060 with 12Gb.

Mostly LORAs this time around, but also one IP-Adapter.

Straight from Ulthar Labs Inc., the new Cat Face Identity IP-Adapter. This is a special kind of plug-in that plugs your cat’s face into each image generation. It’s not just a paste-in, it ‘knows’ what cat faces look like and adapts the source-face to the context. So you could have your favourite domestic kittee go on a trip to Ulthar, etc.

A 20s Silent Movie LORA, trained on scans of good-quality 1920s publicity-stills rather than old scratched celluloid frames, by the look of it. User feedback is good.

Frankenstein Tech – World Morph LORA. A ‘world morph’ LORA that makes everything you prompt for look like it’s either an interior of Frankie’s lab or else came from it. Also good for images of His Monster.

Frankenstein Tech might be combined with also-new Gothic Interiors – v1.1 LORA.

Some more newly-released Lovecraft-adjacent add-ons, which I spotted on CivitAI. For download and local use as style-guides with AI image-generation models derived from Stable Diffusion 1.5. All free, as is the software needed to generate the images (ComfyUI, InvokeAI, etc).

* Declassified Documents LORA.

* Another Limbo Style LORA, based on the visual style of the seminal videogame called Limbo.

* Back-side Light LORA for a creepy or film-noir portrait in darkness.

At low levels, ‘Back-side Light’ could be combined with the new Dystopian City LORA.

I spotted a few more new Lovecraftian-adjacent style-guidance plugins (LORAs), free for the free Stable Diffusion 1.5.

Transparent Specimen LORA, giving an x-ray creature effect. Perhaps especially useful for those making old 1920s x-ray plates, as part of a RPG game?

World of Green Mist LORA, looking very suitable for Innsmouth pictures, though you may want the Sponge tool in Photoshop to tone down the end results. It could be used with the also-new Creature in Myst LORA.

Starry night, for generating an Earth landscape with a starry night sky and moon.

Another pick of the recent free plugins for local/offline use with the free Stable Diffusion 1.5 AI image generator.

* Style of Milton Caniff LORA, able to do 1930s men’s suits and hairstyles in an illustration style. Of obvious relevance for anyone crafting 1930s Lovecraft-era artwork.

* VintageBW LORA for a vintage b&w photography look, and it could be combined with the also-new DarkLight – RimLight lighting LORA for noir rim lighting.

* DonM – Fossils LORA for generating fossils or polished stylised bas-reliefs.

* Mythicc Lands LORA, an unusual landscape generator that combines the landscape with a ‘geometric portal’ composition. Especially suitable for aiding attempts to emulate 1980s style sci-fi paperback cover-art, I’d suggest.

And an oldie, someone made an EMBEDDING (use like a LORA) for SD 1.5 which gave them a consistent Erich Zann face/head for a storytelling project. Still waiting for the consistent-character LORA for Lovecraft himself.

In the meanwhile, Centerflex model + the wasmoebius embedding has a fair idea of what Lovecraft looked like, though you may need to use negatives (Auden, Lincoln, Keaton) and stretch the resulting image up by 70 pixels to get the ‘long face’…

My pick of three recent free Lovecraft-adjacent ‘style plugins’ (LORAs), for use with the free AI image-generator Stable Diffusion 1.5…

Style of Edward Gorey LORA.

Neon Pop Haeckel LORA, Haeckel being an influence on Lovecraft. Though for laying out a “sane” non-religious view of the universe, rather than his deep-sea expedition imagery. Joshi calls him a… “central influence on [Lovecraft’s] metaphysical thought of the time”.

Pulpy Comics Style Inkifier LORA, for a neo-pulp modern comic-book style.

Newly-spotted AI image generation plugins, of possible Lovecraftian interest. For use with Stable Diffusion 1.5 models running locally in InvokeAI, ComfyUI or similar.

Carcosa City.

Ivan Aivazovsky style, for atmospheric sea-scape images (just add monsters). Could be used with the new Cinematic Tools – Volumetric Lighting for backdrops needing more eerie depth-haze/fog.

A few creature plugins of note. Auraplanktons for slides of lit deep-sea specimens.

Swamp Mysteries generates ‘somewhat toony but still mysterious’ swamp monsters in swamps.

And today there’s Black Cat. Seemingly meant to be an eldritch mystery moggie so it looks a bit rough, more Lovecraft’s battered ‘Old Man’ cat than kute kitten. Still, all the more suitable for a desperate messenger from Ulthar, etc.

For steampunk and 1920s RPG makers, Style of patent drawing may interest.