工作流 /
图像 /
MOSS-Image 12GB — text to image + layout boxes + edit LoRA + MoGe control, 1 MP, four workflows

MOSS-Image 12GB — text to image + layout boxes + edit LoRA + MoGe control, 1 MP, four workflows
MOSS-Image 12GB — text to image + layout boxes + edit LoRA + MoGe control, 1 MP, four workflows
☕ One coffee = one happy creator: https://buymeacoffee.com/the_frizzy1MOSS-Image on a 12 GB card, four ways: text to image, layout with bounding boxes, editing one picture, and a new picture in the shape of your photo. F…
图像
5 个节点 2 次使用 The_frizzy1 53 KB
需要的权重
这个工作流引用了 2 个权重文件,其中 0 个能在本站模型库里找到。
The_frizzy1_moss-image-t2i-12gb_v1.0.0.json
模型库暂未收录
The_frizzy1_moss-image-12gb_all-four-workflows_v1.0.0.zip
模型库暂未收录
节点构成
MarkdownNote
× 2
SaveImage
× 1
3301b744-2960-415a-83b0-fe8a89d99fdb
× 1
ResolutionSelector
× 1
官方标注的模型
flux2-vae_small.safetensorsmoss_image_sft_int8_convrot.safetensorsqwen3.5_9b_w4a8_gptq.safetensors
工作流 JSON
和上面「下载工作流 JSON」是同一份文件。
1{
2 "id": "e8f30439-267f-45de-b260-db0d114b2c85",
3 "revision": 0,
4 "last_node_id": 27,
5 "last_link_id": 55,
6 "nodes": [
7 {
8 "id": 9,
9 "type": "SaveImage",
10 "pos": [
11 840,
12 0
13 ],
14 "size": [
15 570,
16 700
17 ],
18 "flags": [],
19 "order": 5,
20 "mode": 0,
21 "inputs": [
22 {
23 "name": "images",
24 "type": "IMAGE",
25 "link": 22
26 }
27 ],
28 "outputs": [
29 {
30 "name": "images",
31 "type": "IMAGE",
32 "links": null
33 }
34 ],
35 "properties": {
36 "Node name for S&R": "SaveImage"
37 },
38 "widgets_values": [
39 "MOSS_t2i"
40 ],
41 "widgets_values_named": {
42 "filename_prefix": "moss_image_t2i"
43 },
44 "title": "Result"
45 },
46 {
47 "id": 18,
48 "type": "3301b744-2960-415a-83b0-fe8a89d99fdb",
49 "pos": [
50 360,
51 0
52 ],
53 "size": [
54 430,
55 680
56 ],
57 "flags": [],
58 "order": 4,
59 "mode": 0,
60 "inputs": [
61 {
62 "label": "positive_prompt",
63 "name": "value",
64 "type": "STRING",
65 "widget": {
66 "name": "value"
67 },
68 "link": null
69 },
70 {
71 "label": "enable_prompt_enhancer",
72 "name": "switch",
73 "type": "BOOLEAN",
74 "widget": {
75 "name": "switch"
76 },
77 "link": null
78 },
79 {
80 "label": "prompt_enhancer_seed",
81 "name": "sampling_mode.seed",
82 "type": "INT",
83 "widget": {
84 "name": "sampling_mode.seed"
85 },
86 "link": null
87 },
88 {
89 "label": "negative_prompt",
90 "name": "text",
91 "type": "STRING",
92 "widget": {
93 "name": "text"
94 },
95 "link": null
96 },
97 {
98 "name": "width",
99 "type": "INT",
100 "widget": {
101 "name": "width"
102 },
103 "link": 40
104 },
105 {
106 "name": "height",
107 "type": "INT",
108 "widget": {
109 "name": "height"
110 },
111 "link": 41
112 },
113 {
114 "label": "SLG",
115 "name": "switch_1",
116 "type": "BOOLEAN",
117 "widget": {
118 "name": "switch_1"
119 },
120 "link": null
121 },
122 {
123 "label": "unet_model",
124 "name": "unet_name",
125 "type": "COMBO",
126 "widget": {
127 "name": "unet_name"
128 },
129 "link": null
130 },
131 {
132 "label": "clip_model",
133 "name": "clip_name",
134 "type": "COMBO",
135 "widget": {
136 "name": "clip_name"
137 },
138 "link": null
139 },
140 {
141 "label": "vae_model",
142 "name": "vae_name",
143 "type": "COMBO",
144 "widget": {
145 "name": "vae_name"
146 },
147 "link": null
148 },
149 {
150 "label": "ref_image (for llm)",
151 "name": "image",
152 "shape": 7,
153 "type": "IMAGE",
154 "link": null
155 }
156 ],
157 "outputs": [
158 {
159 "name": "IMAGE",
160 "type": "IMAGE",
161 "links": [
162 22
163 ]
164 }
165 ],
166 "properties": {
167 "previewExposures": []
168 },
169 "widgets_values": [
170 "Cinematic night street photograph: a model in an exaggerated balloon-sleeve white wool coat walking toward camera under a huge illuminated theatre marquee, the marquee bulbs spelling \"MOSS\" in tall capital letters, wet asphalt with long reflections, warm tungsten glow against cold blue night, 35mm film grain, low angle",
171 true,
172 0,
173 "nsfw, worst quality, bad anatomy, watermark, jpeg artifact",
174 1024,
175 1024,
176 438829726369282,
177 25,
178 2.5,
179 "er_sde",
180 "simple",
181 false,
182 "moss_image_sft_int8_convrot.safetensors",
183 "qwen3.5_9b_w4a8_gptq.safetensors",
184 "flux2-vae_small.safetensors"
185 ],
186 "widgets_values_named": {
187 "value": "Cinematic night street photograph: a model in an exaggerated balloon-sleeve white wool coat walking toward camera under a huge illuminated theatre marquee, the marquee bulbs spelling \"MOSS\" in tall capital letters, wet asphalt with long reflections, warm tungsten glow against cold blue night, 35mm film grain, low angle",
188 "switch": true,
189 "sampling_mode.seed": 0,
190 "text": "nsfw, worst quality, bad anatomy, watermark, jpeg artifact",
191 "width": 1024,
192 "height": 1024,
193 "seed": 438829726369282,
194 "steps": 25,
195 "cfg": 2.5,
196 "sampler_name": "er_sde",
197 "scheduler": "simple",
198 "switch_1": false,
199 "unet_name": "Cinematic night street photograph: a model in an exaggerated balloon-sleeve white wool coat walking toward camera under a huge illuminated theatre marquee, the marquee bulbs spelling \"MOSS\" in tall capital letters, wet asphalt with long reflections, warm tungsten glow against cold blue night, 35mm film grain, low angle",
200 "clip_name": "qwen3.5_9b_w4a8_gptq.safetensors",
201 "vae_name": "flux2-vae_small.safetensors"
202 },
203 "title": "MOSS-Image · text to image",
204 "color": "#223",
205 "bgcolor": "#335"
206 },
207 {
208 "id": 19,
209 "type": "ResolutionSelector",
210 "pos": [
211 10,
212 350
213 ],
214 "size": [
215 280,
216 190
217 ],
218 "flags": [],
219 "order": 0,
220 "mode": 0,
221 "showAdvanced": true,
222 "inputs": [],
223 "outputs": [
224 {
225 "name": "width",
226 "type": "INT",
227 "links": [
228 40
229 ]
230 },
231 {
232 "name": "height",
233 "type": "INT",
234 "links": [
235 41
236 ]
237 }
238 ],
239 "properties": {
240 "Node name for S&R": "ResolutionSelector"
241 },
242 "widgets_values": [
243 "1:1 (Square)",
244 1,
245 16
246 ],
247 "widgets_values_named": {
248 "aspect_ratio": "1:1 (Square)",
249 "megapixels": 1,
250 "multiple": 16
251 }
252 },
253 {
254 "id": 26,
255 "type": "MarkdownNote",
256 "pos": [
257 -1770,
258 0
259 ],
260 "size": [
261 500,
262 520
263 ],
264 "flags": [],
265 "order": 0,
266 "mode": 0,
267 "inputs": [],
268 "outputs": [],
269 "title": "How to use",
270 "properties": [],
271 "widgets_values": [
272 "## How to use - tested on RTX 3060 12 GB\n\n1. Write what you want in the prompt box. Short is fine: the prompt enhancer (the same Qwen3.5 9B that is the text encoder) writes the long caption MOSS likes. You see it in the preview box.\n2. Pick the size in **Resolution**. 1 MP is the default.\n3. Press **Run**. 25 steps, CFG 2.5, er_sde.\n\nRTX 3060, 1 MP, new prompt (enhancer + image): 55 s · same prompt, new seed: 43 s. First run after starting ComfyUI: 1 min 16 s. Peak VRAM 11.5 GB.\n\nDouble-click the main node to see every step."
273 ],
274 "color": "#432",
275 "bgcolor": "#653"
276 },
277 {
278 "id": 27,
279 "type": "MarkdownNote",
280 "pos": [
281 -1770,
282 560
283 ],
284 "size": [
285 500,
286 560
287 ],
288 "flags": [],
289 "order": 0,
290 "mode": 0,
291 "inputs": [],
292 "outputs": [],
293 "title": "Downloads",
294 "properties": [],
295 "widgets_values": [
296 "## Downloads - put each file in the ComfyUI folder shown\n\nClick a file name to download it. Put it in the folder shown (inside `ComfyUI/models/`).\n\n| For | Folder | File (click = download) | Size |\n|---|---|---|---|\n| all | `diffusion_models/` | [moss_image_sft_int8_convrot.safetensors](https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/diffusion_models/moss_image_sft_int8_convrot.safetensors) | 5.41 GB |\n| all | `text_encoders/` | [qwen3.5_9b_w4a8_gptq.safetensors](https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/text_encoders/qwen3.5_9b_w4a8_gptq.safetensors) | 7.34 GB |\n| all | `vae/` | [flux2-vae_small.safetensors](https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/vae/flux2-vae_small.safetensors) | 0.25 GB |\n| edit | `loras/` | [moss_image_edit_lora_rank_64_bf16.safetensors](https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/loras/moss_image_edit_lora_rank_64_bf16.safetensors) | 0.30 GB |\n| control | `loras/` | [moss_image_color_depth_control.safetensors](https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/loras/moss_image_color_depth_control.safetensors) | 0.31 GB |\n| control | `geometry_estimation/` | [moge_3_vitl_fp16.safetensors](https://huggingface.co/Comfy-Org/MoGe/resolve/main/geometry_estimation/moge_3_vitl_fp16.safetensors) | 0.74 GB |\n\nTotal: 14.3 GB. _Tested on: RTX 3060 12 GB · Ryzen 5 5600X · 32 GB RAM · ComfyUI 0.43.1 · sage attention._\n\nNeeds ComfyUI **0.43 or newer**, no custom nodes.\n\nLicense: MOSS-Image: Apache-2.0. Workflow: MIT."
297 ],
298 "color": "#432",
299 "bgcolor": "#653"
300 }
301 ],
302 "links": [
303 [
304 22,
305 18,
306 0,
307 9,
308 0,
309 "IMAGE"
310 ],
311 [
312 40,
313 19,
314 0,
315 18,
316 4,
317 "INT"
318 ],
319 [
320 41,
321 19,
322 1,
323 18,
324 5,
325 "INT"
326 ]
327 ],
328 "groups": [],
329 "definitions": {
330 "subgraphs": [
331 {
332 "id": "3301b744-2960-415a-83b0-fe8a89d99fdb",
333 "version": 1,
334 "state": {
335 "lastGroupId": 5,
336 "lastNodeId": 25,
337 "lastLinkId": 55,
338 "lastRerouteId": 0
339 },
340 "revision": 0,
341 "config": [],
342 "name": "Text to Image (MOSS-Image)",
343 "inputNode": {
344 "id": -10,
345 "bounding": [
346 -1060,
347 -160,
348 191.44921875,
349 368
350 ]
351 },
352 "outputNode": {
353 "id": -20,
354 "bounding": [
355 1630,
356 -10,
357 128,
358 68
359 ]
360 },
361 "inputs": [
362 {
363 "id": "35437e31-565b-4d42-bc06-d95c5929e7d8",
364 "name": "value",
365 "type": "STRING",
366 "linkIds": [
367 25
368 ],
369 "label": "positive_prompt",
370 "pos": [
371 -892.55078125,
372 -136
373 ]
374 },
375 {
376 "id": "853a4192-ebce-42de-bdff-77c36b59cc1a",
377 "name": "switch",
378 "type": "BOOLEAN",
379 "linkIds": [
380 26
381 ],
382 "label": "enable_prompt_enhancer",
383 "pos": [
384 -892.55078125,
385 -116
386 ]
387 },
388 {
389 "id": "cdc1cc70-bfa9-4ec2-84d7-a99f03c6ec6e",
390 "name": "sampling_mode.seed",
391 "type": "INT",
392 "linkIds": [
393 55
394 ],
395 "label": "prompt_enhancer_seed",
396 "pos": [
397 -892.55078125,
398 -96
399 ]
400 },
401 {
402 "id": "c11f1876-7b42-4679-a736-8deaa8f21fd7",
403 "name": "text",
404 "type": "STRING",
405 "linkIds": [
406 43
407 ],
408 "label": "negative_prompt",
409 "pos": [
410 -892.55078125,
411 -76
412 ]
413 },
414 {
415 "id": "1514590b-7a78-450b-82b4-cfbe7820a9db",
416 "name": "width",
417 "type": "INT",
418 "linkIds": [
419 27
420 ],
421 "pos": [
422 -892.55078125,
423 -56
424 ]
425 },
426 {
427 "id": "f2e18bfe-65b0-4d24-9c78-bec0b9573e97",
428 "name": "height",
429 "type": "INT",
430 "linkIds": [
431 28
432 ],
433 "pos": [
434 -892.55078125,
435 -36
436 ]
437 },
438 {
439 "id": "32c26da2-7e60-44bf-851a-72690937b4b4",
440 "name": "seed",
441 "type": "INT",
442 "linkIds": [
443 30
444 ],
445 "pos": [
446 -892.55078125,
447 -16
448 ]
449 },
450 {
451 "id": "dae5d571-ced0-46d6-8a0a-5a71b8ca945a",
452 "name": "steps",
453 "type": "INT",
454 "linkIds": [
455 42
456 ],
457 "pos": [
458 -892.55078125,
459 4
460 ]
461 },
462 {
463 "id": "6c064d6f-9925-4808-83cc-7e3409190ca7",
464 "name": "cfg",
465 "type": "FLOAT",
466 "linkIds": [
467 44
468 ],
469 "pos": [
470 -892.55078125,
471 24
472 ]
473 },
474 {
475 "id": "46e71c72-9f95-4c32-89d2-2dab4b63b284",
476 "name": "sampler_name",
477 "type": "COMBO",
478 "linkIds": [
479 33
480 ],
481 "pos": [
482 -892.55078125,
483 44
484 ]
485 },
486 {
487 "id": "8ba10fc2-8e73-4ece-8f77-27515bbe2698",
488 "name": "scheduler",
489 "type": "COMBO",
490 "linkIds": [
491 34
492 ],
493 "pos": [
494 -892.55078125,
495 64
496 ]
497 },
498 {
499 "id": "71ffd624-a04e-403f-ae06-300f5896f34c",
500 "name": "switch_1",
501 "type": "BOOLEAN",
502 "linkIds": [
503 52
504 ],
505 "label": "SLG",
506 "pos": [
507 -892.55078125,
508 84
509 ]
510 },
511 {
512 "id": "bc8e2627-1b51-4333-9d88-0cb6307f7545",
513 "name": "unet_name",
514 "type": "COMBO",
515 "linkIds": [
516 35
517 ],
518 "label": "unet_model",
519 "pos": [
520 -892.55078125,
521 104
522 ]
523 },
524 {
525 "id": "fe422dee-acbe-4c34-898b-62bb878ddb83",
526 "name": "clip_name",
527 "type": "COMBO",
528 "linkIds": [
529 36
530 ],
531 "label": "clip_model",
532 "pos": [
533 -892.55078125,
534 124
535 ]
536 },
537 {
538 "id": "29a73fdf-7b55-4612-8966-bba0b4cab854",
539 "name": "vae_name",
540 "type": "COMBO",
541 "linkIds": [
542 37
543 ],
544 "label": "vae_model",
545 "pos": [
546 -892.55078125,
547 144
548 ]
549 },
550 {
551 "id": "c8c5a578-f3c4-489e-8fdf-f1cde45fa7e7",
552 "name": "image",
553 "type": "IMAGE",
554 "linkIds": [
555 53
556 ],
557 "label": "ref_image (for llm)",
558 "pos": [
559 -892.55078125,
560 164
561 ]
562 }
563 ],
564 "outputs": [
565 {
566 "id": "fa9fefe4-677e-4f8b-aba5-35459b097c37",
567 "name": "IMAGE",
568 "type": "IMAGE",
569 "linkIds": [
570 6
571 ],
572 "localized_name": "IMAGE",
573 "pos": [
574 1654,
575 14
576 ]
577 }
578 ],
579 "widgets": [],
580 "nodes": [
581 {
582 "id": 17,
583 "type": "PrimitiveStringMultiline",
584 "pos": [
585 -140,
586 -780
587 ],
588 "size": [
589 490,
590 280
591 ],
592 "flags": [],
593 "order": 13,
594 "mode": 0,
595 "inputs": [
596 {
597 "localized_name": "value",
598 "name": "value",
599 "type": "STRING",
600 "widget": {
601 "name": "value"
602 },
603 "link": 25
604 }
605 ],
606 "outputs": [
607 {
608 "localized_name": "STRING",
609 "name": "STRING",
610 "type": "STRING",
611 "links": [
612 15,
613 17
614 ]
615 }
616 ],
617 "title": "Text (User Prompt)",
618 "properties": {
619 "Node name for S&R": "PrimitiveStringMultiline"
620 },
621 "widgets_values": [
622 "Cinematic night street photograph: a model in an exaggerated balloon-sleeve white wool coat walking toward camera under a huge illuminated theatre marquee, the marquee bulbs spelling \"MOSS\" in tall capital letters, wet asphalt with long reflections, warm tungsten glow against cold blue night, anamorphic flare, 35mm film grain, low angle"
623 ],
624 "widgets_values_named": {
625 "value": "Cinematic night street photograph: a model in an exaggerated balloon-sleeve white wool coat walking toward camera under a huge illuminated theatre marquee, the marquee bulbs spelling \"MOSS\" in tall capital letters, wet asphalt with long reflections, warm tungsten glow against cold blue night, anamorphic flare, 35mm film grain, low angle"
626 }
627 },
628 {
629 "id": 14,
630 "type": "PrimitiveStringMultiline",
631 "pos": [
632 -140,
633 -430
634 ],
635 "size": [
636 470,
637 270
638 ],
639 "flags": {
640 "collapsed": false
641 },
642 "order": 0,
643 "mode": 0,
644 "inputs": [],
645 "outputs": [
646 {
647 "localized_name": "STRING",
648 "name": "STRING",
649 "type": "STRING",
650 "links": [
651 13
652 ]
653 }
654 ],
655 "title": "Text (System prompt)",
656 "properties": {
657 "Node name for S&R": "PrimitiveStringMultiline"
658 },
659 "widgets_values": [
660 "You are a prompt engineer for an image generation model. The user sends a short, rough request in Chinese, sometimes with brand names, exact copy to render, or bounding boxes. Expand it into a detailed, self-contained English caption that can go straight into the image model, and make every piece of visible text appear verbatim.\n\nOutput only the caption. No title, preamble, explanation, alternatives, questions, code fences, or restatement of these instructions.\n\nOpening: always start the caption with the medium and the subject, for example \"Anime-style digital illustration of ...\", \"Studio photograph of ...\", \"Mixed-media poster of ...\". Never start with a verb. If the request is an edit, work out the final scene and open on it directly.\n\nHard rule: the caption always describes the finished picture as it looks. It is never an instruction to an editor and never a report of what changed. Never write \"replace\", \"remove\", \"keep\", \"preserve\", \"unchanged\", \"the original\", \"the reference image\", \"the source\" or \"same as\".\n\nPreserve everything the user specifies: number of subjects, identity, action, clothing, style, era, composition, palette, exclusions, and any Chinese copy they quote. Add only plausible visual detail the user left open. Never add new people, plot, or dialogue, never change relationships, and never invent real product specifications, prices, certifications, event details, or real people's names or titles.\n\nBuild the caption from these eight facets. Cover a facet only when the scene has content for it; skip it rather than inventing, and never repeat one to pad length:\n\n1. Subject: appearance, age range, build, hairstyle, each garment and its material, accessories, skin or surface texture.\n2. Action: body orientation, limb positions, weight distribution, what the hands are doing, gaze, expression. The pose must hold in a single frame and never contradict itself.\n3. Space: what sits in the foreground, midground and background, relative positions, occlusion, distances, and what appears at the frame edges.\n4. Setting: architecture and interior elements, furniture, props, signs of use, and the details that place the scene in a time and place (signage, windows, road surface, weather).\n5. Framing: shot size, camera height, viewing angle (level, high, low), perspective strength, where the subject sits in the frame, and the empty space around it.\n6. Light: direction and height of the key light, hardness, color temperature, any practical sources (neon, window light, screens), and how the light falls on skin, fabric, glass, metal and wet ground.\n7. Tone: palette, contrast, grain, medium (digital, film, print, 3D render, painting), texture detail, atmosphere.\n8. Text: see the next section. This is the most common failure point.\n\nText rules:\n- Keep the user's wording verbatim in its original language (Chinese stays Chinese) and write the full string inside double quotes. Never rewrite, translate, shorten, or paste the same copy twice.\n- For each text block state: exact content, carrier (sign, poster, card face, packaging, screen), position and relative size, typeface style and color, and legibility (clearly readable, slightly blurred, partly occluded).\n- When the user gives no copy but the scene clearly needs text (signs, menus, cards, interfaces), author short, scene-appropriate specific wording and give it verbatim. Never write placeholders such as \"headline here\" or \"some text\".\n\nWriting:\n- Flowing natural language, ordered by space or by the viewer's eye, in two to four paragraphs. No tag soup, no \"masterpiece, 8k, best quality\", no empty quality words.\n- Every sentence must add new visible information. Never restate a feature to reach length, and never drop a key detail just to stay brief.\n- Length: aim for 220 to 320 words for a single-subject scene, 320 to 500 for posters, cards, multi-subject scenes and text-heavy layouts, and never fewer than 200 words. A one-word request such as \"1girl\" still gets a full caption of at least 200 words. Do not pad: if you fall short, add the missing visible detail from the eight facets instead of repeating what you already wrote.\n- Describe only what is visible. No tool steps, generation parameters, or coordinate arithmetic.\n- Even when the request contains canvas sizes, resolution, pixels, or wording like \"3:4\" or \"1024x1024\", never repeat any size, ratio or canvas information in the caption.\n\nReference image:\n- Decide whether the supplied image is a content source or a style source. It is a content source when the workflow feeds an image in, or when the user says \"this image\", \"reproduce it\", \"copy it\", \"same as this\", \"only change the background\" or \"only change the copy\". Treat its visible content as the baseline and describe it item by item: subject (hair, hair color, eye color, face, build), every garment (outer layer, inner layer, bottoms, footwear, accessories), props and weapons, pose (body orientation, limb positions, how the hands hold things, weight), framing and how much of the frame the subject fills, background elements and palette, light direction, and art style. Change only what the user asks to change and keep the rest.\n- Describe specifically enough to rebuild, never \"a similar character\" or \"a girl in a similar style\". Correct level of detail: \"a girl with short tousled teal-green hair and long uneven bangs, green eyes, an oversized bright red high-collared jacket with black inner panels over a short black dress, bare legs, chunky black ankle boots with red accents, both hands gripping a massive black blade with neon yellow-green edges, full-body frontal view, white background with red, black and lime-green ink splashes\".\n- When the user asks for the style only (style transfer, \"in this style\", \"same look\"), take the medium, palette, lighting, texture, brushwork and layout logic, and build a new subject and composition instead of copying the original.\n- When the user's text conflicts with the image, the text wins (blue hair means blue hair); everything the text does not mention follows the image.\n- With a very short request such as \"1girl\" plus a reference image, do not improvise: appearance, clothing, palette, background and style all come from the image, and your job is only to write them out as a complete English caption. Do not add accessories, weapons, props or scenery the image does not show.\n- If a detail in the image is unreadable, leave it out rather than guessing. Never guess brands, identities or real names.\n- Text inside the reference image: keep readable wording verbatim under the text rules; write unreadable marks as \"illegible blurred lettering\" instead of inventing content.\n- Never copy watermarks, platform badges or logos the user did not ask for.\n- If this context contains no image at all, do not pretend you saw one: expand the text only, keep every feature the user already gave, and invent no character details.\n- A reference is evidence, not the deliverable. Write the caption as a description of the finished picture, never as an editing instruction to an image editor. Ban edit verbs and comparative pointers: no \"replace\", \"remove\", \"keep\", \"preserve\", \"unchanged\", \"same as\", \"as in the reference\", \"the original\", \"the source\", \"just like\". State the result as plain visible fact instead: the character has short green hair, the background is a night street, the style is loose angular brushwork.\n- Keep the number of subjects consistent with what the user asks for. When they ask to replace the subject, the original subject must not appear in the caption at all.\n\nBounding boxes:\n- Boxes are absolute pixel xyxy (top-left, bottom-right) and are placement instructions only. Use them as given: never normalize, center, reorder, percentage-convert or recompute them.\n- One box only: describe just what is inside that box, one to three sentences, with no coordinates, no numbering, no commentary. Text boxes also state typeface, color, carrier and legibility, with the copy verbatim in double quotes.\n- Several boxes at once: one line per box in the given order, formatted as \"index | English description\", describing only the content inside that box. No coordinates, no JSON, no prose paragraphs.\n- A box marks a placement, not a border, panel, divider or sticker to draw. Overlap and natural occlusion are allowed.\n- With no boxes, write a normal caption and never invent coordinates."
661 ],
662 "widgets_values_named": {
663 "value": "You are a prompt engineer for an image generation model. The user sends a short, rough request in Chinese, sometimes with brand names, exact copy to render, or bounding boxes. Expand it into a detailed, self-contained English caption that can go straight into the image model, and make every piece of visible text appear verbatim.\n\nOutput only the caption. No title, preamble, explanation, alternatives, questions, code fences, or restatement of these instructions.\n\nOpening: always start the caption with the medium and the subject, for example \"Anime-style digital illustration of ...\", \"Studio photograph of ...\", \"Mixed-media poster of ...\". Never start with a verb. If the request is an edit, work out the final scene and open on it directly.\n\nHard rule: the caption always describes the finished picture as it looks. It is never an instruction to an editor and never a report of what changed. Never write \"replace\", \"remove\", \"keep\", \"preserve\", \"unchanged\", \"the original\", \"the reference image\", \"the source\" or \"same as\".\n\nPreserve everything the user specifies: number of subjects, identity, action, clothing, style, era, composition, palette, exclusions, and any Chinese copy they quote. Add only plausible visual detail the user left open. Never add new people, plot, or dialogue, never change relationships, and never invent real product specifications, prices, certifications, event details, or real people's names or titles.\n\nBuild the caption from these eight facets. Cover a facet only when the scene has content for it; skip it rather than inventing, and never repeat one to pad length:\n\n1. Subject: appearance, age range, build, hairstyle, each garment and its material, accessories, skin or surface texture.\n2. Action: body orientation, limb positions, weight distribution, what the hands are doing, gaze, expression. The pose must hold in a single frame and never contradict itself.\n3. Space: what sits in the foreground, midground and background, relative positions, occlusion, distances, and what appears at the frame edges.\n4. Setting: architecture and interior elements, furniture, props, signs of use, and the details that place the scene in a time and place (signage, windows, road surface, weather).\n5. Framing: shot size, camera height, viewing angle (level, high, low), perspective strength, where the subject sits in the frame, and the empty space around it.\n6. Light: direction and height of the key light, hardness, color temperature, any practical sources (neon, window light, screens), and how the light falls on skin, fabric, glass, metal and wet ground.\n7. Tone: palette, contrast, grain, medium (digital, film, print, 3D render, painting), texture detail, atmosphere.\n8. Text: see the next section. This is the most common failure point.\n\nText rules:\n- Keep the user's wording verbatim in its original language (Chinese stays Chinese) and write the full string inside double quotes. Never rewrite, translate, shorten, or paste the same copy twice.\n- For each text block state: exact content, carrier (sign, poster, card face, packaging, screen), position and relative size, typeface style and color, and legibility (clearly readable, slightly blurred, partly occluded).\n- When the user gives no copy but the scene clearly needs text (signs, menus, cards, interfaces), author short, scene-appropriate specific wording and give it verbatim. Never write placeholders such as \"headline here\" or \"some text\".\n\nWriting:\n- Flowing natural language, ordered by space or by the viewer's eye, in two to four paragraphs. No tag soup, no \"masterpiece, 8k, best quality\", no empty quality words.\n- Every sentence must add new visible information. Never restate a feature to reach length, and never drop a key detail just to stay brief.\n- Length: aim for 220 to 320 words for a single-subject scene, 320 to 500 for posters, cards, multi-subject scenes and text-heavy layouts, and never fewer than 200 words. A one-word request such as \"1girl\" still gets a full caption of at least 200 words. Do not pad: if you fall short, add the missing visible detail from the eight facets instead of repeating what you already wrote.\n- Describe only what is visible. No tool steps, generation parameters, or coordinate arithmetic.\n- Even when the request contains canvas sizes, resolution, pixels, or wording like \"3:4\" or \"1024x1024\", never repeat any size, ratio or canvas information in the caption.\n\nReference image:\n- Decide whether the supplied image is a content source or a style source. It is a content source when the workflow feeds an image in, or when the user says \"this image\", \"reproduce it\", \"copy it\", \"same as this\", \"only change the background\" or \"only change the copy\". Treat its visible content as the baseline and describe it item by item: subject (hair, hair color, eye color, face, build), every garment (outer layer, inner layer, bottoms, footwear, accessories), props and weapons, pose (body orientation, limb positions, how the hands hold things, weight), framing and how much of the frame the subject fills, background elements and palette, light direction, and art style. Change only what the user asks to change and keep the rest.\n- Describe specifically enough to rebuild, never \"a similar character\" or \"a girl in a similar style\". Correct level of detail: \"a girl with short tousled teal-green hair and long uneven bangs, green eyes, an oversized bright red high-collared jacket with black inner panels over a short black dress, bare legs, chunky black ankle boots with red accents, both hands gripping a massive black blade with neon yellow-green edges, full-body frontal view, white background with red, black and lime-green ink splashes\".\n- When the user asks for the style only (style transfer, \"in this style\", \"same look\"), take the medium, palette, lighting, texture, brushwork and layout logic, and build a new subject and composition instead of copying the original.\n- When the user's text conflicts with the image, the text wins (blue hair means blue hair); everything the text does not mention follows the image.\n- With a very short request such as \"1girl\" plus a reference image, do not improvise: appearance, clothing, palette, background and style all come from the image, and your job is only to write them out as a complete English caption. Do not add accessories, weapons, props or scenery the image does not show.\n- If a detail in the image is unreadable, leave it out rather than guessing. Never guess brands, identities or real names.\n- Text inside the reference image: keep readable wording verbatim under the text rules; write unreadable marks as \"illegible blurred lettering\" instead of inventing content.\n- Never copy watermarks, platform badges or logos the user did not ask for.\n- If this context contains no image at all, do not pretend you saw one: expand the text only, keep every feature the user already gave, and invent no character details.\n- A reference is evidence, not the deliverable. Write the caption as a description of the finished picture, never as an editing instruction to an image editor. Ban edit verbs and comparative pointers: no \"replace\", \"remove\", \"keep\", \"preserve\", \"unchanged\", \"same as\", \"as in the reference\", \"the original\", \"the source\", \"just like\". State the result as plain visible fact instead: the character has short green hair, the background is a night street, the style is loose angular brushwork.\n- Keep the number of subjects consistent with what the user asks for. When they ask to replace the subject, the original subject must not appear in the caption at all.\n\nBounding boxes:\n- Boxes are absolute pixel xyxy (top-left, bottom-right) and are placement instructions only. Use them as given: never normalize, center, reorder, percentage-convert or recompute them.\n- One box only: describe just what is inside that box, one to three sentences, with no coordinates, no numbering, no commentary. Text boxes also state typeface, color, carrier and legibility, with the copy verbatim in double quotes.\n- Several boxes at once: one line per box in the given order, formatted as \"index | English description\", describing only the content inside that box. No coordinates, no JSON, no prose paragraphs.\n- A box marks a placement, not a border, panel, divider or sticker to draw. Overlap and natural occlusion are allowed.\n- With no boxes, write a normal caption and never invent coordinates."
664 }
665 },
666 {
667 "id": 11,
668 "type": "TextGenerate",
669 "pos": [
670 460,
671 -780
672 ],
673 "size": [
674 400,
675 550
676 ],
677 "flags": [],
678 "order": 9,
679 "mode": 0,
680 "inputs": [
681 {
682 "localized_name": "clip",
683 "name": "clip",
684 "type": "CLIP",
685 "link": 48
686 },
687 {
688 "localized_name": "image",
689 "name": "image",
690 "shape": 7,
691 "type": "IMAGE",
692 "link": 53
693 },
694 {
695 "localized_name": "video",
696 "name": "video",
697 "shape": 7,
698 "type": "IMAGE",
699 "link": null
700 },
701 {
702 "localized_name": "audio",
703 "name": "audio",
704 "shape": 7,
705 "type": "AUDIO",
706 "link": null
707 },
708 {
709 "localized_name": "system_prompt",
710 "name": "system_prompt",
711 "shape": 7,
712 "type": "STRING",
713 "link": 13
714 },
715 {
716 "localized_name": "prompt",
717 "name": "prompt",
718 "type": "STRING",
719 "widget": {
720 "name": "prompt"
721 },
722 "link": 17
723 },
724 {
725 "localized_name": "seed",
726 "name": "sampling_mode.seed",
727 "type": "INT",
728 "widget": {
729 "name": "sampling_mode.seed"
730 },
731 "link": 55
732 }
733 ],
734 "outputs": [
735 {
736 "localized_name": "generated_text",
737 "name": "generated_text",
738 "type": "STRING",
739 "links": [
740 19
741 ]
742 },
743 {
744 "localized_name": "thinking",
745 "name": "thinking",
746 "type": "STRING",
747 "links": null
748 }
749 ],
750 "properties": {
751 "Node name for S&R": "TextGenerate"
752 },
753 "widgets_values": [
754 "",
755 4096,
756 "on",
757 0.7,
758 64,
759 0.95,
760 0.05,
761 1,
762 0,
763 0,
764 false,
765 true,
766 "auto"
767 ],
768 "widgets_values_named": {
769 "prompt": "",
770 "max_length": 4096,
771 "sampling_mode": "on",
772 "sampling_mode.temperature": 0.7,
773 "sampling_mode.top_k": 64,
774 "sampling_mode.top_p": 0.95,
775 "sampling_mode.min_p": 0.05,
776 "sampling_mode.repetition_penalty": 1,
777 "sampling_mode.seed": 0,
778 "sampling_mode.presence_penalty": 0,
779 "thinking": false,
780 "use_default_template": true,
781 "mtp": "auto"
782 }
783 },
784 {
785 "id": 1,
786 "type": "UNETLoader",
787 "pos": [
788 -150,
789 10
790 ],
791 "size": [
792 350,
793 90
794 ],
795 "flags": [],
796 "order": 1,
797 "mode": 0,
798 "inputs": [
799 {
800 "localized_name": "unet_name",
801 "name": "unet_name",
802 "type": "COMBO",
803 "widget": {
804 "name": "unet_name"
805 },
806 "link": 35
807 }
808 ],
809 "outputs": [
810 {
811 "localized_name": "MODEL",
812 "name": "MODEL",
813 "type": "MODEL",
814 "slot_index": 0,
815 "links": [
816 1,
817 51
818 ]
819 }
820 ],
821 "properties": {
822 "Node name for S&R": "UNETLoader",
823 "models": [
824 {
825 "name": "moss_image_sft_int8_convrot.safetensors",
826 "url": "https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/diffusion_models/moss_image_sft_int8_convrot.safetensors",
827 "directory": "diffusion_models"
828 }
829 ]
830 },
831 "widgets_values": [
832 "moss_image_sft_int8_convrot.safetensors",
833 "default"
834 ],
835 "widgets_values_named": {
836 "unet_name": "moss_image_sft_int8_convrot.safetensors",
837 "weight_dtype": "default"
838 },
839 "color": "#223",
840 "bgcolor": "#335"
841 },
842 {
843 "id": 4,
844 "type": "CLIPTextEncode",
845 "pos": [
846 370,
847 30
848 ],
849 "size": [
850 450,
851 520
852 ],
853 "flags": [],
854 "order": 4,
855 "mode": 0,
856 "inputs": [
857 {
858 "localized_name": "clip",
859 "name": "clip",
860 "type": "CLIP",
861 "link": 2
862 },
863 {
864 "localized_name": "text",
865 "name": "text",
866 "type": "STRING",
867 "widget": {
868 "name": "text"
869 },
870 "link": 20
871 }
872 ],
873 "outputs": [
874 {
875 "localized_name": "CONDITIONING",
876 "name": "CONDITIONING",
877 "type": "CONDITIONING",
878 "slot_index": 0,
879 "links": [
880 8
881 ]
882 }
883 ],
884 "title": "Positive Prompt",
885 "properties": {
886 "Node name for S&R": "CLIPTextEncode"
887 },
888 "widgets_values": [
889 "Cinematic night street photograph: a model in an exaggerated balloon-sleeve white wool coat walking toward camera under a huge illuminated theatre marquee, the marquee bulbs spelling \"MOSS\" in tall capital letters, wet asphalt with long reflections, warm tungsten glow against cold blue night, anamorphic flare, 35mm film grain, low angle"
890 ],
891 "widgets_values_named": {
892 "text": "Cinematic night street photograph: a model in an exaggerated balloon-sleeve white wool coat walking toward camera under a huge illuminated theatre marquee, the marquee bulbs spelling \"MOSS\" in tall capital letters, wet asphalt with long reflections, warm tungsten glow against cold blue night, anamorphic flare, 35mm film grain, low angle"
893 },
894 "color": "#232",
895 "bgcolor": "#353"
896 },
897 {
898 "id": 2,
899 "type": "CLIPLoader",
900 "pos": [
901 -150,
902 550
903 ],
904 "size": [
905 350,
906 140
907 ],
908 "flags": [],
909 "order": 2,
910 "mode": 0,
911 "showAdvanced": false,
912 "inputs": [
913 {
914 "localized_name": "clip_name",
915 "name": "clip_name",
916 "type": "COMBO",
917 "widget": {
918 "name": "clip_name"
919 },
920 "link": 36
921 }
922 ],
923 "outputs": [
924 {
925 "localized_name": "CLIP",
926 "name": "CLIP",
927 "type": "CLIP",
928 "slot_index": 0,
929 "links": [
930 2,
931 3,
932 48
933 ]
934 }
935 ],
936 "properties": {
937 "Node name for S&R": "CLIPLoader",
938 "models": [
939 {
940 "name": "qwen3.5_9b_w4a8_gptq.safetensors",
941 "url": "https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/text_encoders/qwen3.5_9b_w4a8_gptq.safetensors",
942 "directory": "text_encoders"
943 }
944 ]
945 },
946 "widgets_values": [
947 "qwen3.5_9b_w4a8_gptq.safetensors",
948 "moss_image",
949 "default"
950 ],
951 "widgets_values_named": {
952 "clip_name": "qwen3.5_9b_w4a8_gptq.safetensors",
953 "type": "moss_image",
954 "device": "default"
955 },
956 "color": "#432",
957 "bgcolor": "#653"
958 },
959 {
960 "id": 7,
961 "type": "KSampler",
962 "pos": [
963 930,
964 0
965 ],
966 "size": [
967 290,
968 290
969 ],
970 "flags": [],
971 "order": 7,
972 "mode": 0,
973 "inputs": [
974 {
975 "localized_name": "model",
976 "name": "model",
977 "type": "MODEL",
978 "link": 49
979 },
980 {
981 "localized_name": "positive",
982 "name": "positive",
983 "type": "CONDITIONING",
984 "link": 8
985 },
986 {
987 "localized_name": "negative",
988 "name": "negative",
989 "type": "CONDITIONING",
990 "link": 9
991 },
992 {
993 "localized_name": "latent_image",
994 "name": "latent_image",
995 "type": "LATENT",
996 "link": 39
997 },
998 {
999 "localized_name": "seed",
1000 "name": "seed",
1001 "type": "INT",
1002 "widget": {
1003 "name": "seed"
1004 },
1005 "link": 30
1006 },
1007 {
1008 "localized_name": "steps",
1009 "name": "steps",
1010 "type": "INT",
1011 "widget": {
1012 "name": "steps"
1013 },
1014 "link": 42
1015 },
1016 {
1017 "localized_name": "cfg",
1018 "name": "cfg",
1019 "type": "FLOAT",
1020 "widget": {
1021 "name": "cfg"
1022 },
1023 "link": 44
1024 },
1025 {
1026 "localized_name": "sampler_name",
1027 "name": "sampler_name",
1028 "type": "COMBO",
1029 "widget": {
1030 "name": "sampler_name"
1031 },
1032 "link": 33
1033 },
1034 {
1035 "localized_name": "scheduler",
1036 "name": "scheduler",
1037 "type": "COMBO",
1038 "widget": {
1039 "name": "scheduler"
1040 },
1041 "link": 34
1042 }
1043 ],
1044 "outputs": [
1045 {
1046 "localized_name": "LATENT",
1047 "name": "LATENT",
1048 "type": "LATENT",
1049 "slot_index": 0,
1050 "links": [
1051 4
1052 ]
1053 }
1054 ],
1055 "properties": {
1056 "Node name for S&R": "KSampler"
1057 },
1058 "widgets_values": [
1059 0,
1060 "randomize",
1061 25,
1062 2.5,
1063 "er_sde",
1064 "simple",
1065 1
1066 ],
1067 "widgets_values_named": {
1068 "seed": 0,
1069 "control_after_generate": "randomize",
1070 "steps": 25,
1071 "cfg": 2.5,
1072 "sampler_name": "er_sde",
1073 "scheduler": "simple",
1074 "denoise": 1
1075 }
1076 },
1077 {
1078 "id": 3,
1079 "type": "VAELoader",
1080 "pos": [
1081 -150,
1082 740
1083 ],
1084 "size": [
1085 350,
1086 60
1087 ],
1088 "flags": [],
1089 "order": 3,
1090 "mode": 0,
1091 "inputs": [
1092 {
1093 "localized_name": "vae_name",
1094 "name": "vae_name",
1095 "type": "COMBO",
1096 "widget": {
1097 "name": "vae_name"
1098 },
1099 "link": 37
1100 }
1101 ],
1102 "outputs": [
1103 {
1104 "localized_name": "VAE",
1105 "name": "VAE",
1106 "type": "VAE",
1107 "slot_index": 0,
1108 "links": [
1109 5
1110 ]
1111 }
1112 ],
1113 "properties": {
1114 "Node name for S&R": "VAELoader",
1115 "models": [
1116 {
1117 "name": "flux2-vae_small.safetensors",
1118 "url": "https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/vae/flux2-vae_small.safetensors",
1119 "directory": "vae"
1120 }
1121 ]
1122 },
1123 "widgets_values": [
1124 "flux2-vae_small.safetensors"
1125 ],
1126 "widgets_values_named": {
1127 "vae_name": "flux2-vae_small.safetensors"
1128 },
1129 "color": "#323",
1130 "bgcolor": "#535"
1131 },
1132 {
1133 "id": 13,
1134 "type": "PreviewAny",
1135 "pos": [
1136 1260,
1137 -780
1138 ],
1139 "size": [
1140 350,
1141 600
1142 ],
1143 "flags": [],
1144 "order": 10,
1145 "mode": 0,
1146 "inputs": [
1147 {
1148 "localized_name": "source",
1149 "name": "source",
1150 "type": "*",
1151 "link": 21
1152 }
1153 ],
1154 "outputs": [
1155 {
1156 "localized_name": "STRING",
1157 "name": "STRING",
1158 "type": "STRING",
1159 "links": [
1160 20
1161 ]
1162 }
1163 ],
1164 "title": "Preview as Text (final prompt)",
1165 "properties": {
1166 "Node name for S&R": "PreviewAny"
1167 },
1168 "widgets_values": [],
1169 "widgets_values_named": []
1170 },
1171 {
1172 "id": 16,
1173 "type": "ComfySwitchNode",
1174 "pos": [
1175 920,
1176 -780
1177 ],
1178 "size": [
1179 280,
1180 110
1181 ],
1182 "flags": [],
1183 "order": 12,
1184 "mode": 0,
1185 "inputs": [
1186 {
1187 "localized_name": "on_false",
1188 "name": "on_false",
1189 "shape": 7,
1190 "type": "STRING",
1191 "link": 15
1192 },
1193 {
1194 "localized_name": "on_true",
1195 "name": "on_true",
1196 "shape": 7,
1197 "type": "STRING",
1198 "link": 19
1199 },
1200 {
1201 "localized_name": "switch",
1202 "name": "switch",
1203 "type": "BOOLEAN",
1204 "widget": {
1205 "name": "switch"
1206 },
1207 "link": 26
1208 }
1209 ],
1210 "outputs": [
1211 {
1212 "localized_name": "output",
1213 "name": "output",
1214 "type": "STRING",
1215 "links": [
1216 21
1217 ]
1218 }
1219 ],
1220 "properties": {
1221 "Node name for S&R": "ComfySwitchNode"
1222 },
1223 "widgets_values": [
1224 true
1225 ],
1226 "widgets_values_named": {
1227 "switch": true
1228 }
1229 },
1230 {
1231 "id": 6,
1232 "type": "EmptyFlux2LatentImage",
1233 "pos": [
1234 -130,
1235 960
1236 ],
1237 "size": [
1238 300,
1239 120
1240 ],
1241 "flags": [],
1242 "order": 6,
1243 "mode": 0,
1244 "inputs": [
1245 {
1246 "localized_name": "width",
1247 "name": "width",
1248 "type": "INT",
1249 "widget": {
1250 "name": "width"
1251 },
1252 "link": 27
1253 },
1254 {
1255 "localized_name": "height",
1256 "name": "height",
1257 "type": "INT",
1258 "widget": {
1259 "name": "height"
1260 },
1261 "link": 28
1262 }
1263 ],
1264 "outputs": [
1265 {
1266 "localized_name": "LATENT",
1267 "name": "LATENT",
1268 "type": "LATENT",
1269 "slot_index": 0,
1270 "links": [
1271 39
1272 ]
1273 }
1274 ],
1275 "properties": {
1276 "Node name for S&R": "EmptyFlux2LatentImage"
1277 },
1278 "widgets_values": [
1279 1024,
1280 1024,
1281 1
1282 ],
1283 "widgets_values_named": {
1284 "width": 1024,
1285 "height": 1024,
1286 "batch_size": 1
1287 },
1288 "color": "#323",
1289 "bgcolor": "#535"
1290 },
1291 {
1292 "id": 5,
1293 "type": "CLIPTextEncode",
1294 "pos": [
1295 370,
1296 610
1297 ],
1298 "size": [
1299 460,
1300 180
1301 ],
1302 "flags": [],
1303 "order": 5,
1304 "mode": 0,
1305 "inputs": [
1306 {
1307 "localized_name": "clip",
1308 "name": "clip",
1309 "type": "CLIP",
1310 "link": 3
1311 },
1312 {
1313 "localized_name": "text",
1314 "name": "text",
1315 "type": "STRING",
1316 "widget": {
1317 "name": "text"
1318 },
1319 "link": 43
1320 }
1321 ],
1322 "outputs": [
1323 {
1324 "localized_name": "CONDITIONING",
1325 "name": "CONDITIONING",
1326 "type": "CONDITIONING",
1327 "slot_index": 0,
1328 "links": [
1329 9
1330 ]
1331 }
1332 ],
1333 "title": "Negative Prompt (empty = reference default)",
1334 "properties": {
1335 "Node name for S&R": "CLIPTextEncode"
1336 },
1337 "widgets_values": [
1338 ""
1339 ],
1340 "widgets_values_named": {
1341 "text": ""
1342 },
1343 "color": "#223",
1344 "bgcolor": "#335"
1345 },
1346 {
1347 "id": 8,
1348 "type": "VAEDecode",
1349 "pos": [
1350 1270,
1351 0
1352 ],
1353 "size": [
1354 230,
1355 70
1356 ],
1357 "flags": [],
1358 "order": 8,
1359 "mode": 0,
1360 "inputs": [
1361 {
1362 "localized_name": "samples",
1363 "name": "samples",
1364 "type": "LATENT",
1365 "link": 4
1366 },
1367 {
1368 "localized_name": "vae",
1369 "name": "vae",
1370 "type": "VAE",
1371 "link": 5
1372 }
1373 ],
1374 "outputs": [
1375 {
1376 "localized_name": "IMAGE",
1377 "name": "IMAGE",
1378 "type": "IMAGE",
1379 "slot_index": 0,
1380 "links": [
1381 6
1382 ]
1383 }
1384 ],
1385 "properties": {
1386 "Node name for S&R": "VAEDecode"
1387 }
1388 },
1389 {
1390 "id": 15,
1391 "type": "SkipLayerGuidanceDiTSimple",
1392 "pos": [
1393 -150,
1394 280
1395 ],
1396 "size": [
1397 350,
1398 220
1399 ],
1400 "flags": [],
1401 "order": 11,
1402 "mode": 0,
1403 "showAdvanced": true,
1404 "inputs": [
1405 {
1406 "localized_name": "model",
1407 "name": "model",
1408 "type": "MODEL",
1409 "link": 1
1410 }
1411 ],
1412 "outputs": [
1413 {
1414 "localized_name": "MODEL",
1415 "name": "MODEL",
1416 "type": "MODEL",
1417 "links": [
1418 50
1419 ]
1420 }
1421 ],
1422 "properties": {
1423 "Node name for S&R": "SkipLayerGuidanceDiTSimple"
1424 },
1425 "widgets_values": [
1426 "",
1427 "22",
1428 0,
1429 0.5
1430 ],
1431 "widgets_values_named": {
1432 "double_layers": "",
1433 "single_layers": "22",
1434 "start_percent": 0,
1435 "end_percent": 0.5
1436 }
1437 },
1438 {
1439 "id": 24,
1440 "type": "ComfySwitchNode",
1441 "pos": [
1442 -110,
1443 140
1444 ],
1445 "size": [
1446 270,
1447 110
1448 ],
1449 "flags": [],
1450 "order": 14,
1451 "mode": 0,
1452 "inputs": [
1453 {
1454 "localized_name": "on_false",
1455 "name": "on_false",
1456 "shape": 7,
1457 "type": "MODEL",
1458 "link": 51
1459 },
1460 {
1461 "localized_name": "on_true",
1462 "name": "on_true",
1463 "shape": 7,
1464 "type": "MODEL",
1465 "link": 50
1466 },
1467 {
1468 "localized_name": "switch",
1469 "name": "switch",
1470 "type": "BOOLEAN",
1471 "widget": {
1472 "name": "switch"
1473 },
1474 "link": 52
1475 }
1476 ],
1477 "outputs": [
1478 {
1479 "localized_name": "output",
1480 "name": "output",
1481 "type": "MODEL",
1482 "links": [
1483 49
1484 ]
1485 }
1486 ],
1487 "properties": {
1488 "Node name for S&R": "ComfySwitchNode"
1489 },
1490 "widgets_values": [
1491 false
1492 ],
1493 "widgets_values_named": {
1494 "switch": false
1495 }
1496 }
1497 ],
1498 "groups": [
1499 {
1500 "id": 1,
1501 "title": "Models",
1502 "bounding": [
1503 -210,
1504 -60,
1505 460,
1506 880
1507 ],
1508 "color": "#3f789e",
1509 "flags": []
1510 },
1511 {
1512 "id": 2,
1513 "title": "Prompt",
1514 "bounding": [
1515 290,
1516 -60,
1517 580,
1518 880
1519 ],
1520 "color": "#3f789e",
1521 "flags": []
1522 },
1523 {
1524 "id": 3,
1525 "title": "Image Size",
1526 "bounding": [
1527 -210,
1528 850,
1529 460,
1530 300
1531 ],
1532 "color": "#3f789e",
1533 "flags": []
1534 },
1535 {
1536 "id": 4,
1537 "title": "Prompt Enhancer",
1538 "bounding": [
1539 -220,
1540 -880,
1541 1900,
1542 840
1543 ],
1544 "color": "#444",
1545 "flags": []
1546 }
1547 ],
1548 "links": [
1549 {
1550 "id": 2,
1551 "origin_id": 2,
1552 "origin_slot": 0,
1553 "target_id": 4,
1554 "target_slot": 0,
1555 "type": "CLIP"
1556 },
1557 {
1558 "id": 20,
1559 "origin_id": 13,
1560 "origin_slot": 0,
1561 "target_id": 4,
1562 "target_slot": 1,
1563 "type": "STRING"
1564 },
1565 {
1566 "id": 1,
1567 "origin_id": 1,
1568 "origin_slot": 0,
1569 "target_id": 15,
1570 "target_slot": 0,
1571 "type": "MODEL"
1572 },
1573 {
1574 "id": 8,
1575 "origin_id": 4,
1576 "origin_slot": 0,
1577 "target_id": 7,
1578 "target_slot": 1,
1579 "type": "CONDITIONING"
1580 },
1581 {
1582 "id": 9,
1583 "origin_id": 5,
1584 "origin_slot": 0,
1585 "target_id": 7,
1586 "target_slot": 2,
1587 "type": "CONDITIONING"
1588 },
1589 {
1590 "id": 21,
1591 "origin_id": 16,
1592 "origin_slot": 0,
1593 "target_id": 13,
1594 "target_slot": 0,
1595 "type": "STRING"
1596 },
1597 {
1598 "id": 15,
1599 "origin_id": 17,
1600 "origin_slot": 0,
1601 "target_id": 16,
1602 "target_slot": 0,
1603 "type": "STRING"
1604 },
1605 {
1606 "id": 19,
1607 "origin_id": 11,
1608 "origin_slot": 0,
1609 "target_id": 16,
1610 "target_slot": 1,
1611 "type": "STRING"
1612 },
1613 {
1614 "id": 3,
1615 "origin_id": 2,
1616 "origin_slot": 0,
1617 "target_id": 5,
1618 "target_slot": 0,
1619 "type": "CLIP"
1620 },
1621 {
1622 "id": 4,
1623 "origin_id": 7,
1624 "origin_slot": 0,
1625 "target_id": 8,
1626 "target_slot": 0,
1627 "type": "LATENT"
1628 },
1629 {
1630 "id": 5,
1631 "origin_id": 3,
1632 "origin_slot": 0,
1633 "target_id": 8,
1634 "target_slot": 1,
1635 "type": "VAE"
1636 },
1637 {
1638 "id": 6,
1639 "origin_id": 8,
1640 "origin_slot": 0,
1641 "target_id": -20,
1642 "target_slot": 0,
1643 "type": "IMAGE"
1644 },
1645 {
1646 "id": 39,
1647 "origin_id": 6,
1648 "origin_slot": 0,
1649 "target_id": 7,
1650 "target_slot": 3,
1651 "type": "LATENT"
1652 },
1653 {
1654 "id": 49,
1655 "origin_id": 24,
1656 "origin_slot": 0,
1657 "target_id": 7,
1658 "target_slot": 0,
1659 "type": "MODEL"
1660 },
1661 {
1662 "id": 50,
1663 "origin_id": 15,
1664 "origin_slot": 0,
1665 "target_id": 24,
1666 "target_slot": 1,
1667 "type": "MODEL"
1668 },
1669 {
1670 "id": 51,
1671 "origin_id": 1,
1672 "origin_slot": 0,
1673 "target_id": 24,
1674 "target_slot": 0,
1675 "type": "MODEL"
1676 },
1677 {
1678 "id": 25,
1679 "origin_id": -10,
1680 "origin_slot": 0,
1681 "target_id": 17,
1682 "target_slot": 0,
1683 "type": "STRING"
1684 },
1685 {
1686 "id": 26,
1687 "origin_id": -10,
1688 "origin_slot": 1,
1689 "target_id": 16,
1690 "target_slot": 2,
1691 "type": "BOOLEAN"
1692 },
1693 {
1694 "id": 43,
1695 "origin_id": -10,
1696 "origin_slot": 3,
1697 "target_id": 5,
1698 "target_slot": 1,
1699 "type": "STRING"
1700 },
1701 {
1702 "id": 27,
1703 "origin_id": -10,
1704 "origin_slot": 4,
1705 "target_id": 6,
1706 "target_slot": 0,
1707 "type": "INT"
1708 },
1709 {
1710 "id": 28,
1711 "origin_id": -10,
1712 "origin_slot": 5,
1713 "target_id": 6,
1714 "target_slot": 1,
1715 "type": "INT"
1716 },
1717 {
1718 "id": 30,
1719 "origin_id": -10,
1720 "origin_slot": 6,
1721 "target_id": 7,
1722 "target_slot": 4,
1723 "type": "INT"
1724 },
1725 {
1726 "id": 42,
1727 "origin_id": -10,
1728 "origin_slot": 7,
1729 "target_id": 7,
1730 "target_slot": 5,
1731 "type": "INT"
1732 },
1733 {
1734 "id": 44,
1735 "origin_id": -10,
1736 "origin_slot": 8,
1737 "target_id": 7,
1738 "target_slot": 6,
1739 "type": "FLOAT"
1740 },
1741 {
1742 "id": 33,
1743 "origin_id": -10,
1744 "origin_slot": 9,
1745 "target_id": 7,
1746 "target_slot": 7,
1747 "type": "COMBO"
1748 },
1749 {
1750 "id": 34,
1751 "origin_id": -10,
1752 "origin_slot": 10,
1753 "target_id": 7,
1754 "target_slot": 8,
1755 "type": "COMBO"
1756 },
1757 {
1758 "id": 52,
1759 "origin_id": -10,
1760 "origin_slot": 11,
1761 "target_id": 24,
1762 "target_slot": 2,
1763 "type": "BOOLEAN"
1764 },
1765 {
1766 "id": 35,
1767 "origin_id": -10,
1768 "origin_slot": 12,
1769 "target_id": 1,
1770 "target_slot": 0,
1771 "type": "COMBO"
1772 },
1773 {
1774 "id": 36,
1775 "origin_id": -10,
1776 "origin_slot": 13,
1777 "target_id": 2,
1778 "target_slot": 0,
1779 "type": "COMBO"
1780 },
1781 {
1782 "id": 37,
1783 "origin_id": -10,
1784 "origin_slot": 14,
1785 "target_id": 3,
1786 "target_slot": 0,
1787 "type": "COMBO"
1788 },
1789 {
1790 "id": 48,
1791 "origin_id": 2,
1792 "origin_slot": 0,
1793 "target_id": 11,
1794 "target_slot": 0,
1795 "type": "CLIP"
1796 },
1797 {
1798 "id": 53,
1799 "origin_id": -10,
1800 "origin_slot": 15,
1801 "target_id": 11,
1802 "target_slot": 1,
1803 "type": "IMAGE"
1804 },
1805 {
1806 "id": 13,
1807 "origin_id": 14,
1808 "origin_slot": 0,
1809 "target_id": 11,
1810 "target_slot": 4,
1811 "type": "STRING"
1812 },
1813 {
1814 "id": 17,
1815 "origin_id": 17,
1816 "origin_slot": 0,
1817 "target_id": 11,
1818 "target_slot": 5,
1819 "type": "STRING"
1820 },
1821 {
1822 "id": 55,
1823 "origin_id": -10,
1824 "origin_slot": 2,
1825 "target_id": 11,
1826 "target_slot": 6,
1827 "type": "INT"
1828 }
1829 ],
1830 "extra": []
1831 }
1832 ]
1833 },
1834 "config": [],
1835 "extra": {
1836 "ds": {
1837 "scale": 0.26442724019097386,
1838 "offset": [
1839 3711.2800949629896,
1840 1839.7684437047312
1841 ]
1842 },
1843 "frontendVersion": "1.55.18"
1844 },
1845 "version": 0.4
1846}