Zhenzhen's AI Radar · MOSS Image · 2026-10-31 Evening Report
This issue features 50 new versions: Strongly Recommended 1 · Recommended 4 · Worth Following 20 · Included Only 25
Editor's Note This issue reviews 51 new MOSS versions, with 1 rated S and 5 rated A. Top pick: Angleview Camera Angle Control LoRA: Teaches MOSS to recognize 19 shooting angles, using "angleview, <angle phrase>" as written, with strength 0.6. The author has disclosed training data and known limitations. A-rated entries include a fully parameterized Iwatomushi style port, Backrooms scene concepts, the Animora semi-realistic anime style, and a Genshin Impact Odette character LoRA with separable clothing. Downgrades: Doubao character LoRAs (commercial brand image, standard character LoRAs), BlackBox, and Fantasy Graphite Sketch ports (missing or copied MOSS parameters) were downgraded from S to B; Test_model_2 shares the same hash as an older model by the same author and was excluded; numerous standard character LoRAs and numbered characters from the Oldies and brats collection were rated C.
Strongly Recommended (1)
[S] Angleview – Camera Angle Control for MOSS Image · v1.0
LoRA · New Model · Author shoukyou · Published 10-31 11:12 · Downloads 99
Camera angle control LoRA: Teaches MOSS to recognize 19 shooting angles (high angle, eye level, low angle, ground-level worm's-eye view × front/side/diagonal rear/back, plus top-down, over-the-shoulder high angle, and raised-hand selfie), solving the issue where MOSS often reverts to a frontal eye-level view despite prompts specifying high angles or rear views.
Maintained at S: Directly addresses MOSS's weakness in camera angles. Angle phrases, strength, tips, and known limitations are clearly documented and actionable; only sampling parameters are missing.
Images
Reasons for Recommendation
- New model: The author notes that MOSS loosely follows camera terminology, often reverting to frontal eye-level views for side-rear high angles, over-the-shoulder views, and very low angles. This LoRA trains each fixed angle phrase individually, serving users who need control over composition and camera position for portraits and scene text-to-image generation.
- The author provides the complete 19 angle phrases (to be used exactly as written), prompt structure, and strength: 0.6, noting that 0.6 and 0.8 yield similar angle control results.
- Usage tips and known limitations are documented: For angles where the face should not be visible, use "her face is not visible"; for low and ground-level angles, add a full-body composition instruction; raised-hand selfies often result in full-body shots from above, and low/ground-level side views sometimes still turn the face toward the camera.
- Training information is public: 207 images (with and without people, mixed groups), each tagged starting with the angle phrase, trained on MOSS Image Base; 9 visible examples, approximately 10 hours after release with 99 downloads and 14 likes (indicating attention only).
How to Use
- Load angleview_moss_v1.safetensors in the MOSS workflow, with strength 0.6 (up to 0.8).
- Start the prompt with "angleview, <angle phrase>." For example, "angleview, from behind and to the side, high angle looking down.", then describe only what is actually visible from that camera position.
- For low/ground-level angles, add a full-body composition instruction; for rear views, use "her face is not visible" instead of describing the face; steps and CFG were not provided by the author.
Risks and Limitations
- This entry has not been tested.
- The angle phrases use the pronoun "her"; it is unclear if this works equally well for male or person-less scenes.
- The author notes that for low angles shot from the front/back, the skirt and hip area may be more exposed than prompted, requiring constraints in the prompt.
- No step or CFG suggestions are provided, nor is it specified if it is compatible with SFT or acceleration LoRAs (training was on Base).
- The license does not allow selling; 2 sensitive-level examples are hidden.
Recommended (4)
[A] iwatomushi style (Anima/Illustrious/Krea2/MOSS/Pony) · MOSS - v1.0
LoRA · Port · Author Mike_832 · Published 10-31 20:29 · Downloads 13
MOSS version of the artist Iwatomushi (iwamushi) style LoRA: Rough, uneven lines, heavy shadows, paper grain, and grayish print colors; default tones are cool and night-oriented. The author notes this is not an official model by the artist.
Downgraded from S to A: The MOSS section parameters, prompt writing, and examples are complete, making this the most thoroughly documented entry in this group; however, it is essentially a port of an artist style imitation, not a new capability, with licensing issues, and the auto-extracted parameters grabbed the wrong base model section.
Images
Reasons for Recommendation
- Port: Previous versions exist for Anima (multiple), Illustrious, Pony, and Krea 2; the MOSS version has a separate section in the full model card, adapted to natural language prompts and trained on MOSS-Image Base. The author states the LoRA primarily adds paper texture, ink lines, and matte colors, while characters and structure come from MOSS. For the closest feel to the artist's hand, use Anima v10.
- Serves users wanting retro, print-grain anime illustrations in MOSS. The author claims rain, snow, neon, and night scenes do not require color words; warm or dark red scenes require specifying a warm light source plus color tags.
- MOSS section parameters are actionable: Prompt template, negative prompts, strength, sampler, scheduler, steps, CFG, shift, and resolution are all provided. Display images retain settings for direct copying.
- 12 visible examples; free, with all site licenses open (generation, renting, selling, selling merges, attribution optional); approximately half an hour after this version's release, 13 downloads, with the entire model accumulating 1342 downloads (indicating attention only).
How to Use
- Load iwamushi_moss_v1.safetensors (36 MB). Start the prompt with "@iwamushi.", followed by full sentences describing the character, action, location, and lighting/mood. No tags, quality words, or style words are needed; strength 1.0 (the author states lowering it only weakens the style).
- The author recommends using MOSS-Image SFT for generation (the LoRA was trained on Base, and tests/display images used SFT int8); er_sde + simple, 25 steps, CFG 3, shift 5, no hires.
- The author's negative prompts are: sensitive content, worst quality, bad anatomy, watermark, jpeg artifact. If speech bubbles or text appear spontaneously, add "speech bubble, text" to the negative prompt and re-roll.
Risks and Limitations
- The auto-extracted parameters (DPM++ 2M SDE, CFG 5, 832×1248, hires 1.5×) are from the Anima v10 section, not the MOSS section. The MOSS section separately recommends skip layer 22, 1024×1536, and no hires, which are only listed in the full model card.
- The MOSS version is step 1800 out of 3000 training steps (the author notes this is not the final step) and was only tested on ComfyUI + MOSS-Image SFT int8.
- The author states the style draws eyes with fewer highlights, and calm expressions may appear tired, requiring emotions like smiles to be specified in the prompt.
- Imitates a specific artist's style, is not official, and training data licensing is not disclosed.
- 7 sensitive-level examples are hidden; this entry has not been tested.
Trigger Words: @iwamushi
[A] Backrooms — Endless Yellow Rooms | MOSS · v1.0
LoRA · New Model · Author 858654128784 · Published 10-31 18:04 · Downloads 29
A concept LoRA for drawing scenes based on the original "Backrooms" internet creepypasta setting: Monochromatic yellow wallpaper, damp beige carpet, buzzing fluorescent lights, and infinitely repeating rooms; it can also draw white-tiled poolrooms. The default look is an overexposed, yellow-green, grainy old digital camera snapshot feel.
Maintained at A: Clear subject, actionable prompt structure and strength, and permissive license; however, sampling parameters are missing, and the subject is narrow, preventing an S rating.
Images
Reasons for Recommendation
- New model: A scene LoRA on MOSS specifically for the Backrooms theme, serving users creating creepypasta, liminal spaces, and horror atmosphere images.
- The author clearly explains the prompt writing: Start with backroomstyle, specify the level (Level 0 yellow rooms, poolrooms, Level 1, Level 2, Level Fun, etc.), then describe walls, floors, and lighting, and finally the appearance, age, position, and action of characters or entities. The author states the most authentic feel comes when the scene is empty, the camera looks old, and characters are small and isolated.
- Strength is provided: 0.8, increasing to 0.9 if the style is weak.
- 7 visible examples; free, with all site licenses open; approximately 3 hours after release, 29 downloads and 9 likes, with a download rate higher than 97% of MOSS models from the last 7 days (indicating attention only).
How to Use
- Load backroomstyle.safetensors (142 MB). Start the prompt with "backroomstyle,", followed by the level name and scene description. When writing characters, specify age as suggested by the author, e.g., "an adult man aged 23".
- Strength 0.8, increasing to 0.9 if necessary; steps, CFG, and sampler were not provided by the author.
Risks and Limitations
- 7 visible examples and 11 sensitive-level examples are hidden, with more hidden than visible.
- The trigger word is not registered in the version info, only mentioned in the description.
- Generation parameters like steps and CFG were not provided by the author; this entry has not been tested.
[A] Animora Style for MOSS-Image · v1
LoRA · New Model · Author Animora · Published 10-31 10:14 · Downloads 41
The author's own semi-realistic anime style LoRA, trained on MOSS using approximately 90 images of different characters. The author claims clean hands, rich lighting, and a unified style across scenes and characters.
Maintained at A: Parameters, trigger words, and mitigation strategies for weaknesses are actionable, and the license is the most permissive; however, with only 1 visible examples, it cannot reach S.
Images
Reasons for Recommendation
- New model: Serves text-to-image generation for character illustrations, fantasy, and daily scenes requiring a unified semi-realistic anime style. The author states it is suitable for stacking with character LoRAs.
- Parameters are fully documented: Strength 1.0, sampler er_sde, scheduler sgm_uniform, 30 steps, CFG 4, 768x1344; trigger word "animora style".
- The author documents weaknesses and countermeasures: Text and signs may garble; back views may appear compressed, in which case lower strength to approximately 0.7 and add "long torso, natural proportions" to the positive prompt.
- Comes with the author's Fire & Forget random scene workflow (another entry in this issue), with the workflow embedded in example images; free, with all licenses open.
How to Use
- Load animorastyle-mossimage.safetensors in the MOSS workflow, write "animora style" in the prompt, with strength 1.0.
- er_sde + sgm_uniform, 30 steps, CFG 4, 768x1344; for back views, lower strength to around 0.7 and add "long torso, natural proportions".
Risks and Limitations
- Only 1 visible examples, with 3 sensitive-level examples hidden, making it difficult to judge style stability from public examples.
- This is the author's personal style, and the source of training materials is not disclosed.
- This entry has not been tested.
Trigger words: animora style
【A】Genshin Odette | Genshin Impact Odette · v1.0
LoRA · New Model · Author Suzuka_naoko · Published 10-31 19:45 · Downloads 11
A MOSS character LoRA for the Genshin Impact character Odette: the dance dress, feather sleeves, different sleeve cuffs and gloves for each hand, headpiece, high heels, and rapier are all separated into individual tags. You can specify exactly which item you want, making it convenient for outfit changes and realistic cosplay-style image generation.
Maintained by Auto A: The breakdown of components, outfit changes, and cosplay instructions are actionable, and parameters are provided; however, it remains a single character, visible examples are too few, and base model recommendations are inconsistent, so it does not reach S-tier.
Images
Reasons for Recommendation
- New model: Compared to standard character LoRAs, it includes component separation, supports Chinese/English trigger words, tags, and natural language, serving text-to-image generation for Genshin Impact fan art, outfit changes, and cosplay photo styles
- The author provided test parameters: SFT base model, 40 steps, CFG 3.5, euler / simple, shift 5, LoRA strength 1.0, 832×1216
- Specific methods for outfit changes and realistic cosplay are provided: For outfit changes, do not write the original clothing components, keep the headpiece, and add the original clothing terms in the negative prompt (do not write the character name in the negative prompt); for cosplay, write "a young East Asian woman cosplaying odette (genshin impact)" in English, followed by camera and lighting details
- Known issues are also noted: In anime-style half-body/full-body images, the finger-hook sleeve cuffs are often simplified and can be fixed via inpainting; blue gloves occasionally turn into long gloves, so add "elbow gloves" to the negative prompt
- Free, with an open license; approximately 1 hours after release, it has 11 downloads (indicating attention level only)
How to Use
- Load odette_v5_s2200.safetensors with a strength of 1.0; write "odette (genshin impact)" or "Odette" in the prompt, as hair and eye colors are already included in the trigger words
- Add component tags as needed: odette dress, odette feather sleeves, odette bridal gauntlet, odette blue glove, odette tiara, odette heels, odette sword (Chinese equivalents: Odette dance dress, Odette feather sleeves, Odette finger-hook sleeve cuffs, etc.)
- Author's parameters: 40 steps, CFG 3.5, euler / simple, shift 5, 832×1216
Risks and Limitations
- This entry has not been tested
- Parameter descriptions are contradictory: The Chinese section states that images from the SFT base model are dirty and recommends pairing with DMD or switching to Base, while the English section and parameters state testing was done on SFT, with DMD not specifically tested
- Only 2 visible examples are available, and 5 sensitive-level examples are hidden
- Game characters are copyrighted by miHoYo; judgment is required for fan works and commercial use
Worth Watching (20)
【B】Doubao | MOSS Image Character LoRA · v1
LoRA · New Model · Author H_ExtremeDream · Published 10-31 14:45 · Downloads 59
A character LoRA for the persona of ByteDance's AI assistant "Doubao": short black hair with red-pink tips, soft and smooth skin texture, positioned by the author for stylized portraits, character portraits, and outfit changes.
Images
【B】AG: Uvall - Fantasy Midjourney · MOSS-Image
LoRA · Ported · Author Azriaelis · Published 10-31 11:00 · Downloads 46
The MOSS version of the Uvall art style LoRA from the "Ars Goetia" series: The author trained on mixed images from Midjourney, NijiJourney, Tumblr, etc., resulting in a fantasy, impasto-style digital illustration aesthetic.
Images
【B】BlackBox - Artist Style · MOSS-Image
LoRA · Ported · Author Azriaelis · Published 10-31 09:30 · Downloads 28
The MOSS version of a style LoRA imitating artist BlackBox; the tags are Korean manhwa, impasto, and semi-realistic, but the model card does not describe the art style itself, only listing recommended parameters for three base models.
Images
【B】Fantasy Graphite Sketch · v10 - MOSS
LoRA · Ported · Author Fatbuns · Published 10-31 10:51 · Downloads 30
Version 10 of the fantasy-themed pencil/graphite sketch style LoRA, this version is the MOSS edition; the model card body still contains the description for the initial Pony version, with the MOSS version having only a brief note on future plans.
Images
【B】MOSS: Fully automated text-to-image/image enhancement/image modification · v1.0
Workflow · New Model · Author NicknickAI · Published 10-31 12:58 · Downloads 56
A ComfyUI workflow JSON for MOSS-Image: Uses a local Qwen vision-language model to expand short prompts or reverse-engineer descriptions from reference images, then generates images with MOSS, finally connecting SeedVR2 for upscaling and before/after comparison.
Images
【B】Grani/grani (Arknights) · v1
LoRA · New Model · Author DSLM · Published 10-31 14:01 · Downloads 16
A character LoRA for the Arknights operator "Grani": gray hair, purple eyes, horse ears, and a ponytail, covering two outfits: the original Elite 0 skin and the "Wonderful Time" skin. Trained on MOSS-Image Base, with SFT recommended for image generation.
Images
【B】LuCiD Visual Regional Painter · 0.1 beta...more notes
Workflow · Ported · Author EB__ · Published 10-31 18:46 · Downloads 5
A placeholder entry for the MOSS version of the ComfyUI plugin LuCiD VRP (Visual Regional Painter): Loads images in ComfyUI, edits layers, and allows dragging boxes on the canvas to write prompts for each region, creating layout prompts for MOSS and regional prompts for Anima.
Images
【B】KTV Room — Drunk Enough to Sing Anything | MOSS · v1.0
LoRA · New Model · Author 858654128784 · Published 10-31 11:49 · Downloads 100
A concept LoRA for Chinese-style KTV private room scenes: leather sofas, fruit platters, beer, and dice cups on a marble-patterned long table, a large TV playing Chinese lyric MVs, a song selection screen, and purple light strips. Can be rendered as ordinary eye-level party photos, or high-angle surveillance footage with white timestamps and room numbers.
Images
【B】1018sus Style | MOSS, Anima · 1018sus for MOSS
LoRA · Ported · Author wideo2015_eon · Published 10-31 12:41 · Downloads 26
The MOSS version of a chibi art style LoRA imitating artist 1018sus: large heads and eyes, slender limbs, and characters often have colorful skin.
Images
【B】MOSS Fire & Forget Wildcards | MOSS-Image Workflow · v1.0
Workflow · New Model · Author Animora · Published 10-31 10:14 · Downloads 24
A ComfyUI random scene workflow for MOSS: Automatically randomly selects location, time/weather, and lighting for each image generation, allowing batch generation of different scenes without writing prompts. Characters or LoRAs are filled into the "Character & Pose" box by the user.
Images
【B】Anime Oil Painting Style · Moss
LoRA · New Model · Author peam3178 · Published 10-31 12:16 · Downloads 60
A soft anime/fantasy oil painting style LoRA; the model card has only a one-sentence description, with no trigger words or parameters.
Images
【B】Foot Spa — It Doesn't Stop at the Foot Rub | MOSS · v1.0
LoRA · New Model · Author 858654128784 · Published 10-31 12:19 · Downloads 55
A concept LoRA for realistic Chinese-style foot massage shop scenes: A row of brown recliners and footstools, wooden foot soaking buckets, towels, tea cups, warm yellow lighting, and steam, used to draw customers and therapists in foot massage shops.
Images
【B】Cinematic Photography · v1.6000
LoRA · New Model · Author aimann · Published 10-31 12:29 · Downloads 50
A realistic style LoRA adding cinematic photography and lighting to MOSS SFT; the description is only a one-sentence title-style summary.
Images
【B】Manga style (Naoki Urasawa) · Moss v1
LoRA · Ported · Author Yofaraway · Published 10-31 18:49 · Downloads 1 · Paid
The MOSS version of a black-and-white manga style LoRA by Naoki Urasawa (training materials mainly from "Monster", "20th Century Boys", and "Pluto"), currently available as a paid download.
Images
【B】AuthentiQ Onesie: Pikachu [Krea2/MOSS] · MOSS_v1.0
LoRA · Ported · Author AuthentiQ · Published 10-31 15:56 · Downloads 30
The MOSS version of a realistic Pikachu onesie (kigurumi-style) clothing LoRA, which can be stacked with character LoRAs to dress characters in Pikachu onesies.
Images
【B】BlackWhiplash style | [MOSS | Krea 2] · moss_1.0
LoRA · Ported · Author NoiseWriggler · Published 10-31 20:53 · Downloads 3
The MOSS version of a Western cartoon/anime style LoRA imitating artist BlackWhiplash, trained on MOSS Base and tested on SFT; zero visible examples.
Images
【B】ks2_iso_gameart · v1
LoRA · New Model · Author mohsensd1373787 · Published 10-31 15:27 · Downloads 28
An isometric game art style LoRA, used for generating isometric game scenes or asset images; the author's description consists of a single sentence: "Testing MOSS for isometric."
Image
【B】bangdream_avemygo_style (Fusing game card art and anime style) · v1.0 for sft
LoRA · Version Update · Author NekoArc17 · Published 10-31 09:11 · Downloads 32
The SFT version of the LoRA for the new generation BanG Dream! (MyGO!!!!!, Ave Mujica, etc.) art style: The author states it fuses two styles—mobile game card art and anime main series visuals—switching between 2D card art style and 2.5D cel-shaded still image feel using mode keywords.
Image
【B】Astrochan (MOSS) · v1
LoRA · New Model · Author Winnougan · Published 10-31 20:03 · Downloads 0 · Paid
A MOSS character LoRA for the author's original mascot "Astrochan" (a girl with pink short hair and a red-and-white spacesuit), currently available as a paid early access.
Image
【B】Graphic novel style for Moss · v1.0 Graphic Novel
LoRA · New Model · Author Liant · Published 10-31 14:31 · Downloads 0 · Paid
A paid LoRA for European adult comic / graphic novel art styles; the author claims it can reproduce various graphic novel styles and works well with prompts like Moebius or 80 era graphic novels.
Image
Included Only (25)
- Chat GPT White Dragon Girl - White Dragon Girl Chat GPT · Moss-White Dragon Girl — LoRA: An adult version character LoRA for the original character "White Dragon Girl" (MOSS); training images were produced by modifying outputs from GPT Image 2.5; a younger version is also available on the same page.
- dracaene sunbringer | Dracaena Sunbringer | ZZZ | Zenless Zone Zero · moss-v1 — LoRA: A MOSS character LoRA for the Zenless Zone Zero character "Dracaena Sunbringer": light gray twin tails, heterochromia, black and blue high-neck short dress, and clover accessories.
- Ebony NordhavenLove · MossImage — LoRA: A MOSS character LoRA for the fictional female character "Ebony NordhavenLove"; the model card consists entirely of character backstory (a glass artisan by the Swedish coast), with no trigger words or usage instructions.
- Braces · v1 — LoRA: A concept LoRA for adding braces (orthodontic appliances) to characters; the model card contains only the single sentence "Braces LoRA for Moss".
- Original Characters | Original Character Compilation LoRA · Luke-MOSS — LoRA: The MOSS version of the newly added character "Luke" (filename lukebennett) in an original character compilation LoRA; the model card only introduces another character, Liu Xiaohu, with no description for Luke himself.
- toma (toma50) style · v1.0 — LoRA: An art style LoRA for the artist toma (toma50); the model card only lists trigger words, strength, and a note saying "can be paired with dmd2 LoRA," without describing the style characteristics.
- Genshin Impact Miliastra Wonderland-Blazing Ambition | Miliastra Wonderland - Blazing Ambition (Cheerleaders) ｜ anima/illust v2 · v1.0 MOSS — LoRA: The MOSS version of a clothing LoRA for the "Blazing Ambition" cheerleader outfit from Genshin Impact's "Miliastra Wonderland": red pleated skirt, crop top, white gloves, rose emblem on the waist, white bow on the back, and white knee-high socks; optional sun hat, whistle, and sneakers.
- Medical Checkup — Medical Examinations | MOSS · v1.0 — LoRA: A concept LoRA for a hospital medical checkup center scene: examination bed with paper, height and weight scale, ultrasound machine, stethoscope and blood pressure monitor, frosted glass windows, and Chinese medical reports, illuminated by cold white fluorescent lights.
- MOSS Illustria Fusion · v1.0 — LoRA: An anime art style LoRA with cel shading; the author states it was "distilled" from the style of a certain Illustrious fine-tuned model, but is unsure which specific model it originated from.
- Daria Vance | Realistic Character LoRA · v1.0 — LoRA: A MOSS character LoRA for the fictional realistic female character "Daria Vance" (dark short hair, freckles); the author claims high facial consistency, suitable for natural candid portraits and lifestyle scenes.
- Rushzilla (3D Style) (Artist Style) [MOSS & IL] · MOSS — LoRA: The MOSS version of a 3D rendering style LoRA imitating the artist Rushzilla; the model card only contains social media links for the author and the artist, with no usage instructions.
- HSRhuangquan · v1 — LoRA: A character LoRA for the Honkai: Star Rail character Huangquan (trigger word HSRhuangquan); the description consists only of the sentence "Just for share, 2 shape".
- Chenbin Style · v1.0 — LoRA: An art style LoRA for the artist Chenbin (chenbin); the card contains only the phrase "a style lora" and the full prompt for one example image.
- HSRkafka · v1 — LoRA: A MOSS character LoRA for the Honkai: Star Rail character Kafka; the model card only says "just for share".
- MARE Style · MOSS — LoRA: The MOSS version of an anime/cartoon style LoRA imitating the artist MARE, also providing four checkpoint saves from training steps 3000, 4000, 5000, and 6000; the description is only one sentence.
- Scorbunny - Valorlynz · Moss — LoRA: A character LoRA for the anthropomorphic "Scorbunny" in the style of artist Valorlynz: a white-furred rabbit-girl, orange eyes, bandage on nose, gold baseball cap, sports bra, and dolphin shorts.
- Oldies and brats Moss image · Brat40 — LoRA: A new character version Brat40 in the character LoRA compilation "Oldies and brats"; this compilation hosts multiple character LoRAs for older women and younger women on the same model page, with each version representing one character.
- raiden_shogun · v1 — LoRA: A MOSS character LoRA for the Genshin Impact character "Raiden Shogun"; the model card only contains the sentence "just for share".
- Oldies and brats Moss image · Brat43 — LoRA: Another new character version Brat43 in the character LoRA compilation "Oldies and brats," uploaded in the same batch as versions like Brat40 from the same compilation, with each version representing one character.
- Oldies and brats Moss image · Brat42 — LoRA: Brat42 in the "Oldies and brats" compilation: The author hosts multiple character LoRAs for older and younger women on the same page, where each version name corresponds to one character; Brat42 is one of them.
- zanniwuwa · v1 — LoRA: A MOSS character LoRA for the Wuthering Waves character "Zani"; the model card only says "just for share," with no registered trigger word.
- yixuan_zzz · v1 — LoRA: A MOSS character LoRA for the Zenless Zone Zero character Yixuan; the model card only says "just for share," with no visible examples.
- Oldies and brats Moss image · Brat44 — LoRA: Brat44, the last one uploaded in this batch for the "Oldies and brats" compilation, is another character version within this multi-character model page; no visible examples.
- Oldies and brats Moss image · Brat41 — LoRA: The "Brat41" entry in the "Oldies and brats" character LoRA compilation: The author publishes multiple older and younger female character LoRAs on the same model page by number; this is one of those characters.
- suisui-wuwa · v1 — LoRA: A MOSS character LoRA for a Wuthering Waves character (name written as suisui); the author only states that there are too few MOSS character LoRAs and trained this one to share, with no trigger word or visible examples.
Data: Civitai public API (base model MOSS Image), inclusion window 10-31 08:31 → 10-31 21:01 (Beijing Time). Recommendations are reviewed by editors; download counts reflect attention only.