MOSS Image's Kodachrome LoRA: 34 MB 4x5 film color transfer
Creator MOSS Image released a 16x-compressed LoKR (only 34 MB) for transferring the color response and sharpness of 4x5 Kodachrome film, trained on 551 scans of 4x5 sheet film.
Creator MOSS Image released a 16x-compressed LoKR (only 34 MB) for transferring the color response and sharpness of 4x5 Kodachrome film, trained on 551 scans of 4x5 sheet film.
The author tried out the ControlNet-LLLite for Anima in ComfyUI, which can use 3D poses to specify character actions and region control to specify the generation range.
AI Canvas v0.10.17 adds "lossless compression" for image nodes; PNG/JPEG keep their original format and pixel dimensions, and a separate asset copy is created only when actual size reduction occurs, with the original image retained and undo supported. This version also fixes a WebKit compatibility issue with large-image preview on Mac and an issue where the input cursor is hard to place after a reference capsule, and improves storage health cleanup and extraction of short-drama characters, scenes, and props. Manual lossless compression is supported only on desktop, with a per-image limit of 32 MiB, 24 million pixels, and 16384 pixels per side.
ComfyUI Comic Creator v1.63.0 updates the OCR tab with new features including Shift multi-select bbox, Ollama VRAM adjustment, decision model empty bbox removal, grayscale preprocessing options, and VLM holistic prompt generation. Custom full-specification mode takes 139.2 seconds, and font differences significantly affect text recognition accuracy.
The author used Animagine XL 4.0 Opt in ComfyUI to generate a Halloween little devil with silver hair, red eyes, short black horns, wearing a black-and-orange outfit and looking up at the camera; the resulting image is 1280×1024.
The creator used ComfyUI and Krea 2 to produce the Indian cyberpunk-themed story "Neo-Mumbai," blending Indian culture, cyberpunk, retro-futurism, and cinematic storytelling. The story revolves around two childhood friends: a smooth-talking drifter and a battle-hardened fixer, who are unexpectedly drawn into a conspiracy involving the city's powerful elite. The author says they have been involved with AI image generation for just over a week, are still learning and experimenting, and welcome criticism and suggestions.
Meta's personal AI agent "Muse" launched in Japan on October 23, the first region outside North America to offer it; users can access it via iOS, Android and muse.ai.
A local Ollama model packaged based on the official moss-image-prompt-enhancer skill, which can expand brief image ideas into detailed prompts for the MOSS-Image text-to-image model, runs offline, and requires no agent or system prompt.
Akane released a set of AI-generated illustrations themed "Ikemen Ojisan?" (Handsome Middle-Aged Man?), with scenes including a café, the street, riding a motorcycle, kneeling on one knee and firing an anti-materiel rifle, leaping between buildings, and an alleyway. The author says they have never been good at drawing "ikemen ojisan" and lack a clear concept of the image.
Several AI companies are hiring human designers to build their brand identities: How&How rebranded Ideogram, Lazarev designed the visual identity for AI video app Elva, Together designed a four-petal logo for AI audit company Clusia, and Meta commissioned type artist Jessica Hische to draw a new logo for its AI platform Muse.
Claude set up a ComfyUI workflow on an M1/16GB Mac; after adjusting launch options, single-step generation went from 10 minutes to about 10 seconds, and they combined processes including a watercolor LoRA, black-and-white manga switching, inpainting, and upscaling.
The author used Animagine XL 4.0 Opt in ComfyUI to generate an example of a female rock singer holding a mic, leaning forward, and belting, at a resolution of 1248×832, Seed 2026101857, Steps 28, CFG 5, euler_ancestral. The accompanying JSON uses only 7 standard nodes and requires no input image, LoRA, ControlNet, or custom nodes; two comparison JSONs are also provided, one removing the forward lean and one removing the singing expression.
Some users found that third-party platforms such as piapi and wiro sell a "seedream 5.0 Pro - uncensored" version capable of generating erotic content, while the same request submitted through ByteDance's official API is rejected. The user wants to obtain the same unrestricted version directly from ByteDance.
A user tested a high token limit on MOSS SFT with a prompt of about 1.3k tokens, combined with a 4-steps turbo LoRA, 4MP resolution, and Euler/simple sampling, to generate a photo containing 12 different people. Each character's appearance, clothing, and props remained consistent, with no color bleeding or duplication.
The author shares hands-on results from training a character LoRA on MOSS Image, saying that using Automagic3 with Sigmoid and the High setting, a learning rate starting at 0.0005 with a minimum learning rate set, and not excluding block 22 produced good results every time.
Aizawa Amu posted 5 illustrations on X from the morning of October 24 to late night on October 25, and freely distributed the corresponding ChatGPT image-generation prompts, covering five scenes: morning, daytime, evening, night, and late night. The prompts require attaching a reference image of an adult character aged 21 or over that the user has the right to use, preserving facial features, hairstyle, hair color, and eye color, and outputting a 1536:1024 landscape single-scene anime-style illustration. The author also suggests changing only one detail, for example swapping the daytime plaid fabric sample for a polka-dot pattern.
A user trained a style LoRA for Qwen-image with ai-toolkit on a Colab 80GB A100, but the output backgrounds were too simple, random people and duplicate characters appeared, and the style drifted. The configuration was rank 16, 1024px, bf16, adamw8bit; a learning rate of 1e-4 for 500 steps had no effect, so it was changed to 3e-4 and trained to 2k steps, and a 12-image dataset still showed no improvement. The user plans to expand to 30–40 images.
Using Animagine XL 4.0 Opt with Control-LoRA Canny rank128 contour guidance, a 1248×832 landscape composition of an elf drawing a bow with Seed 2026101813 was generated in ComfyUI, and the contour PNG and a JSON containing the node configuration were released.
Canva released its first in-house image-generation model, Canva Craft, on October 22, supporting up to 2K resolution and more than 30 aspect ratios, and currently ranking 4th on an image comparison leaderboard. On the same day, KDDI's povo2.0 integrated Nano Banana 2 into povo AI, with image generation free up to 3 times per day.
A collection of 100 AI beauty image-generation prompts for tools such as ChatGPT and Gemini has been released, all in directly copyable JSON format. Each prompt set includes prompt, negative_prompt, and aspect_ratio fields, and provides 1:1, 9:16, and 1280×670 sizes plus background and mood adaptation versions. The themes focus mainly on beauties in spectacular and everyday scenes around the world, such as Argentine salt flats and Kyoto riverside terraces.
The author made a palette custom node for ComfyUI, with about 350 color schemes collected from design, fashion, and animal and plant illustrated-guide websites, selectable with a click. The node replaces color words in prompts via `__color1__` to `__color5__` placeholders; in testing, specifying 3 colors generated themed illustrations such as true red and rose, and sea blue and mint. The definition file is in JSON format and the color schemes can be further expanded.
The author used ChatGPT to generate Halloween-themed images for original characters Popo, Bobo, and Shuka, but the character designs frequently went wrong: Bobo was given hands and feet, Shuka was put in a crooked hat, and inexplicable elements such as "Maple Valley Station" appeared. The author also shows successful cases using watercolor-style prompts that blend watercolor bleeding effects into small objects and the ground, making the white-dress setting for non-human characters work. The author also suggests that when fixing failed images, regenerating from the original reference image has a higher success rate than repeated local edits.
User newsock999 trained a George Barbier style LoRA for Krea2 and has now published it on Civitai. The LoRA draws its style from the haute couture fashion illustrations of the early 20th-century French illustrator; the author reused description text generated by Qwen as the prompt and provides the corresponding prompt and a simple example workflow (pastebin) in the Civitai post.
Google has shut down all three Imagen 4 models in the Gemini API and Vertex AI, previously priced at $0.02 per image for Fast, $0.04 for Standard, and $0.06 for Ultra.
The top three most-downloaded Krea 2 checkpoints on Civitai—Moody Krea 2 Mix, RedCraft, and Muse By Stable Yogi Krea2—all merged LoRAs marked "no derivatives · attribution required." The poster reveals that their own Krea 2 checkpoint is among those using the largest number of LoRAs in merging.
The author used MOSS-Image's JSON bbox specification in ComfyUI to arrange each person individually, testing three sets of scenes with 3, 6, and 10 people; with 10 people, the number of people, arrangement order, and each person's hair color and clothing basically matched the specification, with only one person missing eyes when the LoRA was disabled.
The author has released a set of monolid (single eyelid) prompts for AI image generation, requiring both eyes to be natural monolids with no upper eyelid crease, along with correction prompts for when double eyelids are generated. In testing, ChatGPT's regular Chat vaguely showed monolids, while ChatGPT Work 6Astra instead generated double eyelids; the author says the higher-tier model feels like it has a higher success rate but cannot guarantee it.
The ComfyUI-Moss-Image-Edit-Native node has been updated with support for the akhaliq/MOSS-Image-Edit-HQ LoRA, switchable via edit_format and chinoll's LoRA, each retaining the layout used during training.
Reddit users discuss expectations for future image and video models, covering T2I, I2I, and video generation. The poster says they are satisfied with Krea 2 and only want a better VAE and a slightly larger DiT, and is curious whether open-source models using entirely new architectures (such as the one used by GPT-2.5) will appear.
Visage-Image is a small text-to-image model for face generation. The author built their own dataset, labeled it with models such as Qwen3.5-4B, and also trained a narrow-domain autoencoder; on their dataset its iFID outperforms FLUX.2.
MOSS-Image was tested at native 4MP with Krea 2 prompts, taking about 25 seconds per image and using 16-24GB VRAM on an RTX PRO 6000, based on the official built-in ComfyUI template workflow, with all results being unselected first runs. The model responds well to casual flash-photography texture, subtle skin texture, and stylized proportions, but seed variance is very large—each seed significantly changes the composition, so seeds must be filtered when a specific composition is needed.
A user making posters and infographics in ComfyUI wants to use the native Bounding Box capabilities of multiple models for layout and local inpainting, but is paralyzed by analysis due to too many choices, and asks the community for experience and tips from those who have actually used multiple models or nodes.
The key to creating a rainy-day atmosphere with AI illustration is to manage the direction of light through perspective and to add particle impurities such as raindrops as moderate noise. Using car headlights backlighting a rain curtain as an example, the author points out that the clearer the light band and directionality the better, but the actual effect is only light faintly penetrating from behind. They also emphasize that expression and emotion should come before the weather setting—a smiling face with rain looks eerie.
A user asks the ComfyUI community for help with comic-style redrawing, encountering two core problems: comics contain text, and direct redrawing deforms the text; a single image may contain multiple characters, making it hard to keep their expressions basically consistent with the original. They have tried some methods but can at best change colors and cannot alter the original line art, so they seek workflow or model suggestions.
Higgsfield uses a monthly subscription credit system: Starter is $15 per month with 200 credits, Plus is $49 with 1,000 credits, Ultra is $129 with 3,000 credits, and the team version starts at $79 per seat with 1,000 credits.
Before starting a new round of style LoRA training, select early and late checkpoints and compare them using the same prompts, seeds, base model, resolution, and sampling settings, at both commonly used and lower LoRA strengths, and include a version with LoRA disabled. Then test with prompts for subjects missing from the dataset to check whether only the style was learned without copying objects or compositions from the training images. Finally, validate the best checkpoint and strength with multiple seeds, and record style match, prompt adherence, and artifacts labeled by checkpoint, optimizer step count, and LoRA strength.
The local open-source application ZPix has added lasso local repainting (inpainting) and several familiar selection tools, and supports Photoshop-style shortcuts for full screen, restoring selections, and other operations. This version also provides image-to-image functionality for Krea 2. The demo was recorded on an RTX 3070M 8GB, and the generation time was edited.
A user has encountered a problem generating two-character scenes in ComfyUI: when both characters' faces are visible, their expressions tend to be stiff and lack believable interaction, and the user is unwilling to use a workflow that generates them separately and then composites them. The user is generating static anime-style images rather than video and is seeking techniques to achieve expressive, interactive two-character scenes in a single render.
Claude has begun building a watercolor comic generation workflow for ComfyUI. The user originally only wanted it to organize learning entry points for the three keywords watercolor painting, comics, and art supplies, but it went ahead and assembled the pipeline directly. The author then asked GPT to compile an implementation instruction document, writing character features, paper white space, color layering and bleeding, and post-generation inspection and correction steps into a form that could be handed to Claude. The instruction document is now complete, but whether the workflow runs successfully and the quality of the finished watercolor comics remain to be verified.
Loading MOSS-Image's color-depth-control LoRA in ComfyUI allows the use of depth and color map images to control generation results.