MOSS Image Morning Report 2026-11-22 | Featured: Syokumura - Artist Style

This issue reviewed 44 new MOSS versions, with 1 rated S, 13 rated A, 11 rated B, and 19 rated C; none were excluded. The featured pick is the Syokumura art style: sketchy lines combined with watercolor washes and localized bright accents, ported from the Krea 2 / Anima versions. The author provided specific MOSS trigger words, strength, steps, and sampler settings. Other highlights include an HDR LoRA that adjusts tone without changing subject matter, Ebiblue's black-and-white illustration old style, Megami magazine anime style v3, a MOSS sampler comparison workflow, and several distinct pieces from Dizzle's astronomy-themed batch, including auroras, Pluto, full moon silhouettes, and frost. The auto-rated S Dark Fantasy and Realistic Blend base models were downgraded to A due to missing MOSS-specific parameters or descriptions; the Fortnite character

Zhenzhen's AI Radar · MOSS Image · 2026-11-22 Morning Report

This issue features 34 new versions: Strongly Recommended 1 · Recommended 10 · Worth Watching 10 · Listed Only 13

Editor's Note This issue reviewed 44 new MOSS versions, with 1 rated S, 13 rated A, 11 rated B, and 19 rated C; none were excluded. The featured pick is the Syokumura art style: sketchy lines combined with watercolor washes and localized bright accents, ported from the Krea 2 / Anima versions. The author provided specific MOSS trigger words, strength, steps, and sampler settings. Other highlights include an HDR LoRA that adjusts tone without changing subject matter, Ebiblue's black-and-white illustration old style, Megami magazine anime style v3, a MOSS sampler comparison workflow, and several distinct pieces from Dizzle's astronomy-themed batch, including auroras, Pluto, full moon silhouettes, and frost. The auto-rated S Dark Fantasy and Realistic Blend base models were downgraded to A due to missing MOSS-specific parameters or descriptions; the Fortnite character Bravo was downgraded to B; the AI Babe Pack 4 sub-packages and Marvel character batch ports are rated C.

Strongly Recommended (1)

[S] Syokumura - Artist Style · MOSS-Image

LoRA · Port · Author Azriaelis · Published 11-22 00:00 · Downloads 44

Artist Syokumura's art style LoRA: expressive lines with a sketchy feel, soft watercolor-style washes, localized vivid color accents, and cinematic composition; a MOSS port of the Krea 2 and Anima versions.

Images

Example 1024x1536

Reasons for Recommendation

  • Port: The same page already features Krea 2 Turbo and Anima versions, with the entire model page having 658 downloads; the model card includes a dedicated set of parameters written specifically for MOSS, rather than copying settings from other base models.
  • Suitable for illustrations, concept art, atmospheric scenes, and text-to-image character generation when you want a "pencil sketch + watercolor coloring" texture.
  • 16 visible examples (another 4 sensitive-level images are hidden); published 9 hours ago with 44 downloads; the author states that including the trigger word significantly strengthens the style.

How to Use

  • Trigger word: syokumura (stronger effect when included)
  • MOSS settings: Base model + 8 steps turbo LoRA, sampler euler, scheduler simple, LoRA strength 0.8-1, steps 8-16, CFG 1-3

Risks and Limitations

  • The MOSS version requires the use of a 8 step turbo LoRA; the author did not specify which turbo LoRA to use.
  • Imitates an individual artist's style, with no authorization stated by the author; the license does not allow derivatives, commercial use is limited to image output and on-site rental; version notes are empty.

Trigger Word: syokumura

Recommended (10)

[A] Dark Fantasy illustration [Krea + MOSS + Flux1/2 + ZiT + IL + Pony] · MOSS

LoRA · Port · Author ColonelJay · Published 11-22 06:30 · Downloads 20

Dark fantasy illustration style LoRA: thick black lines, hatching texture, extreme light-dark contrast, and oppressive color palettes of dark red/earthy brown/cold blue, resembling RPG illustrations like "Darkest Dungeon"; a MOSS port from a multi-base model series.

(Editor downgraded to A: The style description is clear, but the author did not provide MOSS-specific parameters.)

Images

Example 896x1120

Reasons for Recommendation

  • Port: The same page already features Krea 2, Flux.1, Flux.2 Klein, Z-Image Turbo, Illustrious, and Pony versions, with the entire model page having 982 downloads; the MOSS version filename indicates round 10.
  • The author clearly breaks down the style into four points (thick lines and hatching, exaggerated grotesque shapes, muted colors, oppressive lighting), suitable for dungeon exploration, dark RPGs, and board game card illustrations.
  • Trigger words are explicit, with two extended prompts provided; the license is permissive, allowing image output, rental, derivatives, and no attribution required.

How to Use

  • Trigger word: D4rk_f4nt4sy_illustr4tion; for natural language base models, you can append "dark fantasy illustration in a hatched style, with thick lines and pronounced textures".
  • The strength values provided by the author are for Flux / Z-Image (1) and Illustrious / Pony (0.8 to 1); the author did not provide strength, steps, or CFG for the MOSS version.

Risks and Limitations

  • 6 visible examples, with another 13 sensitive-level images hidden.
  • The MOSS version notes only state "MOSS image version," with no description of differences from other base model versions.

Trigger Words: D4rk_f4nt4sy_illustr4tion, D4rk_f4nt4sy_illustr4tion, dark fantasy illustration in a hatched style, with thick lines and pronounced textures., D4rk_f4nt4sy_illustr4tion, digital art, dark fantasy illustration, hatched style, fantasy setting, rough sketch, heavy linework,

[A] kkw_hdr_moss · v1.0

LoRA · New Model · Author devilkkw · Published 11-21 21:47 · Downloads 61

An HDR effect LoRA for MOSS SFT, requiring no trigger word: low strength is used to enhance details in natural photography, while high strength produces a distinct HDR style.

Running the same realistic photo through 0, low strength, and 1 allows you to directly see the boundary between "detail enhancement" and "HDR" usage. (Editor downgraded to A: No examples for low-strength usage, and examples rely on unnamed custom nodes.)

Images

Example 2416x1808

Reasons for Recommendation

  • New Model: A utility LoRA that adjusts tone without changing subject matter, stackable on any realistic photography prompt; the author provided two strength ranges corresponding to "detail enhancement" and "HDR".
  • Trained using OmniLora on moss-sft_int8_convrot; 8 visible examples, published 11 hours ago with 61 downloads.
  • All examples were generated using the realistic base model IntoRealism at strength 1.

How to Use

  • No trigger word required; strength 0.05-0.4 is for natural photography and brightening details, while higher strength up to 1 produces HDR effects (the author's original text says "05 to 1," the starting point appears to be a typo).
  • The author did not provide steps, CFG, or sampler settings; examples used a custom node downloadable in ComfyUI, but the node name was not specified by the author.

Risks and Limitations

  • All examples show HDR effects at strength 1; there are no examples for the low-strength "detail enhancement" usage.
  • The author is unsure if it works on MOSS Base; the version name is marked as starter, the model card is very short; the license does not allow derivatives.

[A] Ebiblue - Gesugao - Old Style (2016-2020) - MOSS LORA · v1.0

LoRA · New Model · Author Konan · Published 11-22 04:16 · Downloads 31

A "gesugao" LoRA based on artist Ebiblue's old style from 2016-2020: high-contrast black-and-white manga illustrations, with only a pair of glowing eyes as the sole color, trained directly on MOSS.

(Editor downgraded to A: Prompts are well-provided, but parameters like steps are missing, and it imitates an artist's style.)

Images

Example 1280x1840

Reasons for Recommendation

  • New Model: A style template from author Konan, but this version provides three directly combinable prompts—an opening sentence, an eye color accent sentence, and a style enhancement sentence.
  • Suitable for black-and-white manga illustrations and villain/dark character portraits; the "only eyes have color" localized accent effect is controlled via prompts, with customizable colors.
  • Strength range specified as 0.8-1.3; 6 visible examples, published less than 5 hours ago with 31 downloads and 6 likes.

How to Use

  • Opening: gesugao, High-contrast monochrome illustration of + [scene content]; Eyes: Her eyes are glowing bright XXXX as the only selective color, framed by heavy dark shadows across her face. (Replace XXXX with a color).
  • Style Enhancement: Dark atmospheric background, clean ink lines, deep shading, greyscale style; strength 0.8-1.3, author did not provide steps or CFG.

Risks and Limitations

  • Imitates an individual artist's style, with no authorization stated by the author; another 3 sensitive-level examples are hidden.
  • No textual description of style characteristics; the latter half of the model card consists of donation and commission advertisements.

Trigger Word: Her eyes are glowing bright XXXX as the only selective color, framed by heavy dark shadows across her face.

[A] Dizzle's Northern Lights · v1

LoRA · New Model · Author Maridizzle · Published 11-22 00:51 · Downloads 34

An aurora-themed concept LoRA: splashes ribbons of green, purple, and electric blue auroras onto any subject matter, making characters semi-transparent and fabrics flow like light; part of Dizzle's space/nature series.

(Editor downgraded to A: One of several mass-produced theme LoRAs from the same author in this batch, providing only strength and trigger words; retains A rating as it shows the most distinct differences among the group.)

Images

Example 1024x1024

Reasons for Recommendation

  • New Model: A "theme filter" not limited to specific subjects; the author cites harpists, giant cats chasing people, and attic bedrooms as examples that can adopt the aurora texture, suitable for fantasy illustrations, posters, and atmospheric images.
  • The same author released multiple models in this issue; this one differs most clearly from several nebula-themed ones: primary colors are green, purple, and blue, focusing on semi-transparent glass-like characters and flowing light fabrics.
  • Trigger word and strength are specified; 6 visible examples, among the highest count in the series.

How to Use

  • Trigger word: dz_northern, strength 0.9 to 1.0; the author states no additional magic words are needed.
  • The author did not provide steps, CFG, or sampler settings.

Risks and Limitations

  • The model card is promotional copy with no training or parameter details; multiple models from the same author in this batch share a template, so stability of individual models is unknown.
  • Another 2 sensitive-level examples are hidden.

Trigger Word: dz_northern

[A] Dizzle's Pluto · v1

LoRA · New Model · Author Maridizzle · Published 11-21 21:54 · Downloads 25

A Pluto-themed concept LoRA: lavender twilight, glowing ice terraces, and low-hanging moons; any subject matter transforms into a soft, pastel, dreamy space style; part of Dizzle's space series.

(Editor downgraded to A: One of several mass-produced theme LoRAs from the same author in this batch, providing only strength and trigger words; retains A rating as the pastel route differs significantly from the nebula version.)

Images

Example 1536x1024

Reasons for Recommendation

  • New Model: Unlike the nebula version in the same series, this one takes a low-saturation pastel route; the author claims even monsters appear gentle, suitable for gentle fantasy illustrations, picture books, and posters.
  • The author cites couples at desks, surfing, and women dragging clock towers as applicable scenarios; the image will consistently feature several moons hanging in the sky.
  • Trigger word and strength are specified; 7 visible examples.

How to Use

  • Trigger word: dz_pluto, strength 0.9 to 1.0.
  • The author did not provide steps, CFG, or sampler settings.

Risks and Limitations

  • The author states that several moons will be added to the top of the image; it is not explained whether they can be removed if unwanted.
  • The model card is promotional copy with no training or parameter details; multiple models from the same author in this batch share a template; another 2 sensitive-level examples are hidden.

Trigger Word: dz_pluto

[A] Ancient Style Xianxia · Ancient Style Xianxia MOSS Version

LoRA · Port · Author nimengchuan520 · Published 11-22 00:07 · Downloads 57

A LoRA for ancient Chinese fantasy (xianxia) style. The author of the Krea 2 version describes it as “MJ-like xianxia style”; this is a ported version for moss_image_sft_int8.

Images

Example 1072x1920

Reasons to Recommend

  • Port: The same page already has a Krea 2 version, with the entire model page having 751 downloads and 80 likes; the MOSS version explicitly corresponds to the moss_image_sft_int8 model.
  • Complete MOSS parameters: Strength 0.8-1, 30 steps, negative text strength 3.5; no trigger word required.
  • Serves ancient Chinese fantasy characters, scene illustrations, and posters; 6 vertical examples are visible, all non-sensitive level.

How to Use

  • Base model moss_image_sft_int8, strength 0.8-1, steps 30, negative text strength 3.5 (author's original words, likely referring to CFG).
  • The author did not register a trigger word; simply write the content of the ancient Chinese fantasy scene.

Risks and Limitations

  • The style description is only one sentence: “MJ-like xianxia style”; the differences between the MOSS version and the Krea 2 version are not specified.
  • The license does not allow derivatives, requires attribution, and commercial use is limited to on-site rental.

【A】Nickel Nike Style [MOSS] · MOSS v1.0

LoRA · New Model · Author Imagine__That · Published 11-21 23:28 · Downloads 56

Nickel Nike style LoRA: The author describes it as “elegant modern ligne claire,” referring to the Franco-Belgian comic style with uniform thin lines and flat color blocks.

Images

Example 1024x1536

Reasons to Recommend

  • New model: Trained directly for MOSS, with the trigger word and strength range for standalone use clearly specified.
  • Serves comic storyboards, picture books, ligne claire posters, and character illustrations.
  • Published about 9 hours ago with 56 downloads and 7 likes.

How to Use

  • Trigger word: nickel nike; strength 0.8-1.00 when used alone.
  • The author did not provide strength, steps, or CFG for use in combination with other LoRAs.

Risks and Limitations

  • The description is only two lines; only 3 examples are visible, with another 5 sensitive-level images hidden.
  • If Nickel Nike is a living artist, the author did not specify authorization; the license does not allow commercial use of generated images, limited to on-site rental and rental.

Trigger Word: nickel nike

【A】Dizzle's Snowflakes · v1

LoRA · New Model · Author Maridizzle · Published 11-22 08:00 · Downloads 10

Frost texture concept LoRA: Covers any subject (skin, fruit, ruins, knots, marble) with a layer of ice fern-like white frost and snowflakes. The author claims the original colors of the subject remain unchanged.

Running one image with the LoRA on and one off for the same subject and seed can reveal whether it only adds frost or also changes the shape.

Images

Example 1024x1024

Reasons to Recommend

  • New model: One of the MOSS concept LoRAs released by Maridizzle in the same batch. Other models in the batch are space-themed (aurora, Pluto, nebula), making this the only frost texture option, distinct from the rest.
  • Serves winter-themed posters, icy-textured still life and product images, and close-ups of “frost-covered” characters; examples given by the author range from pumpkins and marble to human skin.
  • Complete parameters: Trigger word dz_snow, recommended strength 1.0; 3 visible examples, free, license allows derivatives without attribution.

How to Use

  • Write the trigger word dz_snow in the prompt, followed by the subject to be frosted; strength 1.0.
  • The author did not provide steps, CFG, or sampler settings.

Risks and Limitations

  • Published about 1 hours ago, with only 3 visible examples and another 2 sensitive-level images hidden.
  • “Colors unchanged, only frost added” is the author's claim; performance on subjects outside the examples is unknown.

Trigger Word: dz_snow

【A】MOSS-Image Sampler Compare: best sampler and scheduler in one run · v1

Workflow · New Model · Author Stable_Yogi · Published 11-21 21:57 · Downloads 71

MOSS sampler comparison workflow: Runs multiple “sampler + scheduler” combinations in one go using the same prompt and seed, outputting an annotated comparison grid and a time-cost table sorted from fastest to slowest.

Images

Example 1280x720

Reasons to Recommend

  • New workflow (not new weights): Solves the choice of “which sampler and scheduler are fastest while still rendering text correctly.” The author suggests writing text intended for the image in quotes within the prompt for comparison.
  • Complete dependency list: Author's SY Nodes custom nodes (Sampler Sweep, Decode, Image Grid, Save Result), ComfyUI 0.43 or newer, qwen3.5_9b_w4a8_gptq text encoder (7.3 GB, type set to moss_image), flux2-vae_small (250 MB).
  • Includes a YouTube tutorial video; published 11 hours ago with 71 downloads and 12 likes (indicating attention level only).

How to Use

  • Drag the .json file into ComfyUI; in the Sampler Sweep SY node, write one “sampler + scheduler” combination per line. Fix the seed and run; the grid is saved to ComfyUI/output.
  • The author's own Realism and Muse MOSS base models use 8 steps and CFG 1; the time cost for the first group includes model loading, so ignore it or run twice.

Risks and Limitations

  • Depends on the author's custom node package; nodes and models must be installed and verified manually. This tool has not been run.
  • Only 1 visible examples; license does not allow derivatives.

【A】Dizzles - The Moon · v1

LoRA · New Model · Author Maridizzle · Published 11-21 23:21 · Downloads 29

Giant full moon silhouette composition LoRA: The background is filled by a full moon with the rest darkened, turning foreground characters or animals into silhouettes. Adding “a black silhouette” to the prompt switches between pure black silhouettes and dramatic lighting effects.

Images

Example 1536x1024

Reasons to Recommend

  • New model: The most distinct one in the author's space series batch—others change color tones and materials, while this one fixes the layout of “giant moon + foreground silhouette.”
  • The author provides toggle words: Adding “a black silhouette” produces full black silhouettes; omitting it retains details but creates moody dramatic lighting.
  • Serves posters, book covers, and gothic or fantasy silhouette scenes; examples given by the author include kneeling women, mermaids on reefs, people praying in candlelit cathedrals, and herds of zebras by wooden boardwalks.

How to Use

  • Trigger word: dz_lunar, strength 0.9–1.0.
  • To get pure black silhouettes, add “a black silhouette” to the prompt; the author did not provide steps, CFG, or sampler settings.

Risks and Limitations

  • Only 2 visible examples, with another 3 sensitive-level images hidden.
  • The author wrote the trigger word as “dz_lunar (confirm)” on the card, indicating they hadn't finalized it themselves.

Trigger Word: dz_lunar

Worth Watching (10)

【B】Bravo - Fortnite · Bravo_Moss

LoRA · Port · Author AthosAI · Published 11-22 08:44 · Downloads 6

Character LoRA for Bravo from Fortnite (dark skin, short beard, undercut pompadour), with item-by-item prompts for green and navy blue uniforms; a MOSS port of the Krea 2, Anima, and Illustrious versions.

Images

Example 832x1216

【B】NekoboshiYuki style · NekoboshiYuki_MOSS_v1.1

LoRA · Version Update · Author yyumt · Published 11-21 21:23 · Downloads 65

New MOSS version v1.1 of the NekoboshiYuki artist style LoRA: The author claims it fixes the issue in the previous version where characters often grew cat ears.

Images

Example 1024x1536

【B】Dizzle's Eagle Nebula · v1

LoRA · New Model · Author Maridizzle · Published 11-22 04:15 · Downloads 21

Eagle Nebula themed concept LoRA: Orange-purple nebula gas fills the scene, blue star clusters twinkle, hair turns into light strands, and stones become clouds; one of Dizzle's space series.

Images

Example 1024x1024

【B】Dizzle's Pillars of Creation · v1

LoRA · New Model · Author Maridizzle · Published 11-22 00:06 · Downloads 31

Pillars of Creation themed concept LoRA: Orange, purple, and cyan nebulae envelop everything, making skin, fabric, and hay glow into light strands; one of Dizzle's space series, with a recommended strength higher than others in the series.

Images

Example 1024x1024

【B】poderdohokuto (conto) - MOSS Image v1 · v3.0

LoRA · Version Update · Author inpaint · Published 11-22 02:11 · Downloads 38

v3.0 of the poderdohokuto (conto) artist style LoRA: Reduces the tendency for “characters always having open mouths” compared to the previous version; the file remains a 7 MB FP8.

Images

Example 832x1216

【B】Horror Vibe - Moss · v1.0

LoRA · New Model · Author EvilPeterPan · Published 11-22 02:07 · Downloads 63

“Horror Vibe” style LoRA. The model card only says “This is my Horror Vibe LoRA,” so the specific type of horror style can only be judged from the examples.

Images

Example 1088x1920

【B】Tom Fischback-Style-Ilustrious+NoobAi+Anima+Krea2+Moss · MossyTom

LoRA · Port · Author WhiteCaliberFox · Published 11-22 01:22 · Downloads 30

Style LoRA for Tom Fischbach, creator of the webcomic Twokinds (Western comic style with beast-ear and furry characters); a MOSS port of the Illustrious, NoobAI, Anima, and Krea 2 versions.

Images

Example 1152x1536

【B】Super Extreme Expressions · v1

LoRA · New Model · Author Eloc_Starwind · Published 11-22 08:03 · Downloads 17

Exaggerated expression concept LoRA: The author's description is only “focus on extreme expressions,” used to push character expressions to highly exaggerated levels.

Images

Example 1024x1024

【B】MOSS Image: 50-step vs 8-step DMD vs 4-step Turbo (ComfyUI) · v1.0

Workflow · New Model · Author AIKSK · Published 11-22 05:50 · Downloads 8

A ComfyUI comparison workflow that runs the same MOSS image generation task side-by-side across three tiers: 50-step SFT as a quality reference, 8-step DMD as the daily default, and 4-step Turbo as the speed tier.

Image

Diagram 1672x941

【B】MOSS Image: image cleaning workflow (ComfyUI) · v1.0

Workflow · New Model · Author AIKSK · Published 11-22 05:50 · Downloads 8

A ComfyUI workflow for cleaning old MOSS images: it uses MOSS to repaint rough, compressed originals, preserving composition while reconstructing surface texture, before passing them to FLUX 3 Klein for upscaling.

Image

Diagram 1086x1448

Included Only (13)


Data: Civitai Public API (base model MOSS Image), inclusion window 11-21 20:31 → 11-22 09:01 (Beijing Time). Recommendation reasons are editor-reviewed; download counts only reflect attention.

Civitai

Explore the latest AI Radar updates