工作流 /
图像 /
Fibo 2.0 Edit Turbo 12GB — int8 Image Editing in 4 steps + Qwen3-VL Prompt Enhancer

Fibo 2.0 Edit Turbo 12GB — int8 Image Editing in 4 steps + Qwen3-VL Prompt Enhancer
Fibo 2.0 Edit Turbo 12GB — int8 Image Editing in 4 steps + Qwen3-VL Prompt Enhancer
☕ One coffee = one happy creator: https://buymeacoffee.com/the_frizzy1No video on this workflow yet — the one above is how to use my workflows, if you are new to them.Fibo 2.0 Edit Turbo on a 12 GB card: say what the pic…
图像
18 个节点 3 次使用 The_frizzy1 54 KB
需要的权重
这个工作流引用了 1 个权重文件,其中 0 个能在本站模型库里找到。
The_frizzy1_fibo2-edit-turbo-12gb_v1.0.0.json
模型库暂未收录
节点构成
CLIPLoader
× 2
PrimitiveStringMultiline
× 2
MarkdownNote
× 2
UNETLoader
× 1
VAELoader
× 1
TextEncodeFibo2
× 1
KSampler
× 1
VAEDecode
× 1
SaveImage
× 1
LoadImage
× 1
TextGenerate
× 1
ComfySwitchNode
× 1
ImageCompare
× 1
PreviewAny
× 1
BatchImagesNode
× 1
官方标注的模型
fibo2_edit_turbo_int8_convrot.safetensorsflux2-vae.safetensorsqwen3vl_4b_bf16.safetensorsqwen3vl_4b_int8_convrot.safetensorsqwen3vl_8b_int8_convrot.safetensors
工作流 JSON
和上面「下载工作流 JSON」是同一份文件。
1{
2 "id": "c968c8f4-5788-4d2a-aea6-c4849ebd7b5d",
3 "revision": 0,
4 "last_node_id": 31,
5 "last_link_id": 39,
6 "nodes": [
7 {
8 "id": 1,
9 "type": "UNETLoader",
10 "pos": [
11 -100,
12 90
13 ],
14 "size": [
15 410,
16 90
17 ],
18 "flags": [],
19 "order": 0,
20 "mode": 0,
21 "inputs": [],
22 "outputs": [
23 {
24 "color_off": null,
25 "color_on": null,
26 "dir": null,
27 "label": null,
28 "locked": null,
29 "name": "MODEL",
30 "nameLocked": null,
31 "removable": null,
32 "type": "MODEL",
33 "pos": null,
34 "slot_index": null,
35 "links": [
36 4
37 ]
38 }
39 ],
40 "properties": {
41 "Node name for S&R": "UNETLoader",
42 "models": [
43 {
44 "name": "fibo2_edit_turbo_int8_convrot.safetensors",
45 "url": "https://huggingface.co/Comfy-Org/Fibo-2.0/resolve/main/diffusion_models/fibo2_edit_turbo_int8_convrot.safetensors",
46 "hash": "1f6fa7c499191f405467123f655f11762661eaa5bd479d1da8f161398ef72c3b",
47 "hash_type": "SHA256",
48 "directory": "diffusion_models"
49 }
50 ]
51 },
52 "widgets_values": [
53 "fibo2_edit_turbo_int8_convrot.safetensors",
54 "default"
55 ],
56 "widgets_values_named": {
57 "unet_name": "fibo2_edit_turbo_int8_convrot.safetensors",
58 "weight_dtype": "default"
59 }
60 },
61 {
62 "id": 2,
63 "type": "CLIPLoader",
64 "pos": [
65 -100,
66 230
67 ],
68 "size": [
69 410,
70 110
71 ],
72 "flags": [],
73 "order": 1,
74 "mode": 0,
75 "inputs": [],
76 "outputs": [
77 {
78 "color_off": null,
79 "color_on": null,
80 "dir": null,
81 "label": null,
82 "locked": null,
83 "name": "CLIP",
84 "nameLocked": null,
85 "removable": null,
86 "type": "CLIP",
87 "pos": null,
88 "slot_index": null,
89 "links": [
90 1
91 ]
92 }
93 ],
94 "properties": {
95 "Node name for S&R": "CLIPLoader",
96 "models": [
97 {
98 "name": "qwen3vl_4b_int8_convrot.safetensors",
99 "url": "https://huggingface.co/Comfy-Org/Fibo-2.0/resolve/main/text_encoders/qwen3vl_4b_bf16.safetensors",
100 "hash": "36f3ff447ef59201722e8f9ce6020c9819fdcfba6aa2608c4e09b1c0ce114e34",
101 "hash_type": "SHA256",
102 "directory": "text_encoders"
103 }
104 ]
105 },
106 "widgets_values": [
107 "qwen3vl_4b_int8_convrot.safetensors",
108 "fibo2",
109 "default"
110 ],
111 "widgets_values_named": {
112 "clip_name": "qwen3vl_4b_int8_convrot.safetensors",
113 "type": "fibo2",
114 "device": "default"
115 }
116 },
117 {
118 "id": 3,
119 "type": "VAELoader",
120 "pos": [
121 -100,
122 400
123 ],
124 "size": [
125 410,
126 60
127 ],
128 "flags": [],
129 "order": 2,
130 "mode": 0,
131 "inputs": [],
132 "outputs": [
133 {
134 "color_off": null,
135 "color_on": null,
136 "dir": null,
137 "label": null,
138 "locked": null,
139 "name": "VAE",
140 "nameLocked": null,
141 "removable": null,
142 "type": "VAE",
143 "pos": null,
144 "slot_index": null,
145 "links": [
146 2,
147 9
148 ]
149 }
150 ],
151 "properties": {
152 "Node name for S&R": "VAELoader",
153 "models": [
154 {
155 "name": "flux2-vae.safetensors",
156 "url": "https://huggingface.co/Comfy-Org/Fibo-2.0/resolve/main/vae/flux2-vae.safetensors",
157 "hash": "868fe7b343cc8f3a19dbcfcafbc3d5f888802be3f89bd81b65b3621a066ce8f3",
158 "hash_type": "SHA256",
159 "directory": "vae"
160 }
161 ]
162 },
163 "widgets_values": [
164 "flux2-vae.safetensors"
165 ],
166 "widgets_values_named": {
167 "vae_name": "flux2-vae.safetensors"
168 }
169 },
170 {
171 "id": 4,
172 "type": "TextEncodeFibo2",
173 "pos": [
174 430,
175 130
176 ],
177 "size": [
178 390,
179 340
180 ],
181 "flags": [],
182 "order": 14,
183 "mode": 0,
184 "showAdvanced": true,
185 "inputs": [
186 {
187 "color_off": null,
188 "color_on": null,
189 "dir": null,
190 "label": null,
191 "locked": null,
192 "name": "clip",
193 "nameLocked": null,
194 "removable": null,
195 "type": "CLIP",
196 "pos": null,
197 "link": 1
198 },
199 {
200 "color_off": null,
201 "color_on": null,
202 "dir": null,
203 "label": null,
204 "locked": null,
205 "name": "images.image_1",
206 "nameLocked": null,
207 "removable": null,
208 "shape": 7,
209 "type": "IMAGE",
210 "pos": null,
211 "link": 3
212 },
213 {
214 "name": "images.image_2",
215 "shape": 7,
216 "type": "IMAGE",
217 "link": null
218 },
219 {
220 "name": "images.image_3",
221 "shape": 7,
222 "type": "IMAGE",
223 "link": null
224 },
225 {
226 "name": "images.image_4",
227 "shape": 7,
228 "type": "IMAGE",
229 "link": null
230 },
231 {
232 "color_off": null,
233 "color_on": null,
234 "dir": null,
235 "label": null,
236 "locked": null,
237 "name": "vae",
238 "nameLocked": null,
239 "removable": null,
240 "shape": 7,
241 "type": "VAE",
242 "pos": null,
243 "link": 2
244 },
245 {
246 "color_off": null,
247 "color_on": null,
248 "dir": null,
249 "label": null,
250 "locked": null,
251 "name": "prompt",
252 "nameLocked": null,
253 "removable": null,
254 "type": "STRING",
255 "widget": {
256 "name": "prompt"
257 },
258 "link": 33
259 }
260 ],
261 "outputs": [
262 {
263 "color_off": null,
264 "color_on": null,
265 "dir": null,
266 "label": null,
267 "locked": null,
268 "name": "positive",
269 "nameLocked": null,
270 "removable": null,
271 "type": "CONDITIONING",
272 "pos": null,
273 "slot_index": null,
274 "links": [
275 5
276 ]
277 },
278 {
279 "color_off": null,
280 "color_on": null,
281 "dir": null,
282 "label": null,
283 "locked": null,
284 "name": "negative",
285 "nameLocked": null,
286 "removable": null,
287 "type": "CONDITIONING",
288 "pos": null,
289 "slot_index": null,
290 "links": [
291 6
292 ]
293 },
294 {
295 "color_off": null,
296 "color_on": null,
297 "dir": null,
298 "label": null,
299 "locked": null,
300 "name": "latent",
301 "nameLocked": null,
302 "removable": null,
303 "type": "LATENT",
304 "pos": null,
305 "slot_index": null,
306 "links": [
307 7
308 ]
309 }
310 ],
311 "properties": {
312 "Node name for S&R": "TextEncodeFibo2"
313 },
314 "widgets_values": [
315 "Turn the coat into black leather with fine silver floral embroidery.",
316 "",
317 0,
318 0,
319 1
320 ],
321 "widgets_values_named": {
322 "prompt": "Turn the coat into black leather with fine silver floral embroidery.",
323 "negative_prompt": "",
324 "width": 0,
325 "height": 0,
326 "batch_size": 1
327 },
328 "color": "#322",
329 "bgcolor": "#533"
330 },
331 {
332 "id": 5,
333 "type": "KSampler",
334 "pos": [
335 930,
336 100
337 ],
338 "size": [
339 320,
340 500
341 ],
342 "flags": [],
343 "order": 15,
344 "mode": 0,
345 "inputs": [
346 {
347 "color_off": null,
348 "color_on": null,
349 "dir": null,
350 "label": null,
351 "locked": null,
352 "name": "model",
353 "nameLocked": null,
354 "removable": null,
355 "type": "MODEL",
356 "pos": null,
357 "link": 4
358 },
359 {
360 "color_off": null,
361 "color_on": null,
362 "dir": null,
363 "label": null,
364 "locked": null,
365 "name": "positive",
366 "nameLocked": null,
367 "removable": null,
368 "type": "CONDITIONING",
369 "pos": null,
370 "link": 5
371 },
372 {
373 "color_off": null,
374 "color_on": null,
375 "dir": null,
376 "label": null,
377 "locked": null,
378 "name": "negative",
379 "nameLocked": null,
380 "removable": null,
381 "type": "CONDITIONING",
382 "pos": null,
383 "link": 6
384 },
385 {
386 "color_off": null,
387 "color_on": null,
388 "dir": null,
389 "label": null,
390 "locked": null,
391 "name": "latent_image",
392 "nameLocked": null,
393 "removable": null,
394 "type": "LATENT",
395 "pos": null,
396 "link": 7
397 }
398 ],
399 "outputs": [
400 {
401 "color_off": null,
402 "color_on": null,
403 "dir": null,
404 "label": null,
405 "locked": null,
406 "name": "LATENT",
407 "nameLocked": null,
408 "removable": null,
409 "type": "LATENT",
410 "pos": null,
411 "slot_index": null,
412 "links": [
413 8
414 ]
415 }
416 ],
417 "properties": {
418 "Node name for S&R": "KSampler"
419 },
420 "widgets_values": [
421 1234,
422 "fixed",
423 4,
424 1,
425 "euler",
426 "simple",
427 1
428 ],
429 "widgets_values_named": {
430 "seed": 1234,
431 "control_after_generate": "fixed",
432 "steps": 4,
433 "cfg": 1,
434 "sampler_name": "euler",
435 "scheduler": "simple",
436 "denoise": 1
437 }
438 },
439 {
440 "id": 6,
441 "type": "VAEDecode",
442 "pos": [
443 1340,
444 50
445 ],
446 "size": [
447 300,
448 50
449 ],
450 "flags": [],
451 "order": 16,
452 "mode": 0,
453 "inputs": [
454 {
455 "color_off": null,
456 "color_on": null,
457 "dir": null,
458 "label": null,
459 "locked": null,
460 "name": "samples",
461 "nameLocked": null,
462 "removable": null,
463 "type": "LATENT",
464 "pos": null,
465 "link": 8
466 },
467 {
468 "color_off": null,
469 "color_on": null,
470 "dir": null,
471 "label": null,
472 "locked": null,
473 "name": "vae",
474 "nameLocked": null,
475 "removable": null,
476 "type": "VAE",
477 "pos": null,
478 "link": 9
479 }
480 ],
481 "outputs": [
482 {
483 "color_off": null,
484 "color_on": null,
485 "dir": null,
486 "label": null,
487 "locked": null,
488 "name": "IMAGE",
489 "nameLocked": null,
490 "removable": null,
491 "type": "IMAGE",
492 "pos": null,
493 "slot_index": null,
494 "links": [
495 10
496 ]
497 }
498 ],
499 "title": "VAEDecode",
500 "properties": {
501 "Node name for S&R": "VAEDecode"
502 }
503 },
504 {
505 "id": 7,
506 "type": "SaveImage",
507 "pos": [
508 1340,
509 160
510 ],
511 "size": [
512 460,
513 500
514 ],
515 "flags": [],
516 "order": 17,
517 "mode": 0,
518 "inputs": [
519 {
520 "color_off": null,
521 "color_on": null,
522 "dir": null,
523 "label": null,
524 "locked": null,
525 "name": "images",
526 "nameLocked": null,
527 "removable": null,
528 "type": "IMAGE",
529 "pos": null,
530 "link": 10
531 }
532 ],
533 "outputs": [
534 {
535 "color_off": null,
536 "color_on": null,
537 "dir": null,
538 "label": null,
539 "locked": null,
540 "name": "images",
541 "nameLocked": null,
542 "removable": null,
543 "type": "IMAGE",
544 "pos": null,
545 "slot_index": null,
546 "links": [
547 30
548 ]
549 }
550 ],
551 "title": "Result",
552 "properties": [],
553 "widgets_values": [
554 "Fibo2_edit"
555 ],
556 "widgets_values_named": {
557 "filename_prefix": "image_fibo2_edit_turbo_pe"
558 }
559 },
560 {
561 "id": 10,
562 "type": "LoadImage",
563 "pos": [
564 450,
565 -650
566 ],
567 "size": [
568 430,
569 530
570 ],
571 "flags": [],
572 "order": 3,
573 "mode": 0,
574 "inputs": [],
575 "outputs": [
576 {
577 "color_off": null,
578 "color_on": null,
579 "dir": null,
580 "label": null,
581 "locked": null,
582 "name": "IMAGE",
583 "nameLocked": null,
584 "removable": null,
585 "type": "IMAGE",
586 "pos": null,
587 "slot_index": null,
588 "links": [
589 3,
590 31,
591 34
592 ]
593 },
594 {
595 "color_off": null,
596 "color_on": null,
597 "dir": null,
598 "label": null,
599 "locked": null,
600 "name": "MASK",
601 "nameLocked": null,
602 "removable": null,
603 "type": "MASK",
604 "pos": null,
605 "slot_index": null,
606 "links": null
607 }
608 ],
609 "title": "Load Image (source)",
610 "properties": {
611 "Node name for S&R": "LoadImage"
612 },
613 "widgets_values": [
614 "auburn_afro_payphone_portrait.png",
615 "image"
616 ],
617 "widgets_values_named": {
618 "image": "auburn_afro_payphone_portrait.png",
619 "upload": "image"
620 },
621 "color": "#322",
622 "bgcolor": "#533"
623 },
624 {
625 "id": 12,
626 "type": "PrimitiveStringMultiline",
627 "pos": [
628 -90,
629 940
630 ],
631 "size": [
632 400,
633 140
634 ],
635 "flags": {
636 "collapsed": false
637 },
638 "order": 4,
639 "mode": 0,
640 "inputs": [],
641 "outputs": [
642 {
643 "color_off": null,
644 "color_on": null,
645 "dir": null,
646 "label": null,
647 "locked": null,
648 "name": "STRING",
649 "nameLocked": null,
650 "removable": null,
651 "type": "STRING",
652 "pos": null,
653 "slot_index": null,
654 "links": [
655 13
656 ]
657 }
658 ],
659 "title": "String (System Prompt VGL)",
660 "properties": {
661 "Node name for S&R": "PrimitiveStringMultiline"
662 },
663 "widgets_values": [
664 "You are a highly skilled Creative Director for Visual Adaptation at a leading Generative AI company.\n\nYour expertise lies in using an existing image as a visual reference to create entirely new scenes. You can deconstruct a reference image to understand its subject, pose, and style, and then reimagine it in a new context based on textual instructions.\n\nYour primary task is to receive a reference image and a textual edit instruction. You will analyze the reference to extract key visual information and then generate a comprehensive JSON object describing a new scene that creatively incorporates the user's instructions.\n\n**FIDELITY RULE:** Unless explicitly requested or implied by the user's edit instruction (e.g., changing the setting requires new lighting), you should preserve all visual properties of the reference image. This includes, but is not limited to: the identity, pose, and appearance of all subjects (people/animals), the existence, location, size, and orientation of all objects, the overall composition, camera angle, lens characteristics, and the style/medium. Only change what is strictly required to fulfill the user's edit.\nE.g. if the user's request is to change the expression of a person in the image, the only change should be the expression, not the pose, composition, background, etc. On the other hand, if the user requested to change the background to a snowy forest, you should change the background, lighting, and mood atmosphere, etc., but keep the subject the same.\nMake sure to acknowledge the changing elements in the relevant \"description\" fields (i.e. \"short_description\", object \"description\", etc.). e.g. if the user requested to change the expression of a person to \"sad\", make sure to write \"the person is sad\" in the \"short_description\".\n\n**Describe the FINAL image only, never the edit (STRICT):** describe changed elements by their resulting STATE, not the action — e.g. `the person looks sad` / `his hands are empty`, NOT `the expression was changed to sad` / `the book was removed`. In every field except `edit_instruction`, do NOT use edit/change verbs or references to the edit (\"changed\", \"added\", \"removed\", \"replaced\", \"now\", \"instead of\", \"has been\", \"originally\", \"the original\"), and never make \"the scene\"/\"the image\"/\"the frame\" the subject of a change verb. Only `edit_instruction` states the change.\n\n**CONTROL-MAP / CONDITIONING INPUTS (act differently):** Sometimes the input image is NOT a normal photograph or illustration but a *conditioning / control map* — e.g. a Canny edge map, a depth map, a pose skeleton, a segmentation or color grid, a low-resolution / tiled image (super-resolution), or a monochrome image to be recolored. You can tell from the instruction (e.g. \"match this edge map\", \"use this depth map to generate ...\", \"use this pose guidance to generate ...\", \"recolor this monochrome image into ...\", \"increase the clarity and resolution\"). In these cases the FIDELITY RULE above is INVERTED for appearance: do NOT preserve or describe the control map's literal look (its edges, grey depth values, skeleton lines, grid colors, blockiness, or lack of color). Instead treat the input purely as STRUCTURAL guidance — outline/geometry (edges), spatial layout and relative distance (depth), body position and framing (pose), regional layout (segmentation/color grid), or coarse composition (low-res) — and describe the NEW, fully-rendered, realistic image the instruction asks for. Take the subject, colors, materials, textures, and style from the instruction's target description (defaulting to a realistic photograph unless it says otherwise), while keeping the structure/geometry/pose that the control map dictates. Every field (`objects`, `background_setting`, `style_medium`, `lighting`, etc.) must describe that intended final image — never the control map itself.\n\nAdhere strictly to the following structure and guidelines:\n\n1. **Inputs:** You will receive a reference image and a textual edit instruction. You will NOT receive a starting JSON.\n2. **Core Logic (Analyze and Synthesize):**\n * **Analyze:** First, deeply analyze the provided reference image. Identify its primary subject(s), their specific poses, expressions, and appearance. Also note the overall composition, lighting style, and artistic medium.\n * **Synthesize:** Next, interpret the textual edit instruction to understand what elements to keep from the reference and what to change. You will then construct a brand new JSON object from scratch that describes the desired final scene. For example, if the instruction is \"the same dog and pose, but at the beach,\" you must describe the dog from the reference image in the `objects` array but create a new `background_setting` for a beach, with appropriate `lighting` and `mood_atmosphere`. **The description of every unchanged element (objects, people, composition, camera settings, style) MUST be derived directly from the reference image.**\n3. **Output:** Your output MUST be ONLY a single, valid JSON object that describes the **new, imagined scene**. Do not describe the original reference image.\n4. **Holistic Consistency:** Ensure the generated JSON is internally consistent. A change in the environment should be reflected logically across multiple fields, such as `background_setting`, `lighting`, `shadows`, and the `short_description`.\n5. **Schema Adherence:** The new JSON object you generate must strictly follow the schema provided below.\n\nUnless the user's prompt explicitly requests a different style (e.g., 'painting', 'cartoon', 'illustration') and if the original image is a photograph, you MUST default to `style_medium: \"photograph\"`. In such cases, `artistic_style` should be \"realistic\" or a similar descriptor.\n\n**Edit Instruction Rewriting Rule:** You should rewrite the user's request into a single instruction in the imperative style. Follow the examples below to guide your rewriting:\n\n| Edit Category | User Input Example | Your Rewritten Output (Imperative Style) |\n| :--- | :--- | :--- |\n| **Modify Style** | \"I want this picture to look like a cartoon.\" | `Turn the image into the cartoon style.` |\n| | \"Could you make it an oil painting?\" | `Change the image style to Oil_Painting.` |\n| **Object Attribute** | \"Make the dog black and white instead of brown.\"| `Change the dog's color to black and white.` |\n| | \"I want the shirt to be a formal white shirt.\"| `Change the subject's shirt to a formal white shirt.` |\n| **Adding Elements** | \"Put a hat on the person's head.\" | `Add a wide-brimmed felt hat to the subject.` |\n| **Removing Objects** | \"Get rid of the book they are holding.\" | `Remove the book from the subject's hands.` |\n| **Replacing Objects** | \"Replace the rose with a sunflower.\" | `Change the rose to a bright yellow sunflower.` |\n| **Lighting/Relight** | \"Change the lighting to be bright and vibrant.\" | `Change the lighting from dark and moody to bright and vibrant.` |\n| **Composition/Camera** | \"I want a wider shot, showing more of the background.\"| `Change the perspective to a wider shot.` |\n| **Text/Label** | \"Change the text 'Happy Anniversary' to 'Hello'.\"| `Change the text \"Happy Anniversary\" to \"Hello\".` |\n| **Quality/Refinement** | \"Can you make the image sharper?\" | `Refine the image to obtain increased clarity and sharpness.` |\n| **Style Reference** | \"Use the style in the image but with a fox.\" | `Use the style in the image to depict a red fox walking.` |\n\nThe JSON object must contain the following keys precisely:\n\n1. `short_description`: (String) A concise summary of the imagined image content, 200 words maximum.\n2. `objects`: (Array of Objects) List a maximum of 5 prominent objects. If the scene implies more than 5, creatively choose the most important ones and describe the rest in the background. For each object, include:\n * `description`: (String) A detailed description of the imagined object, 100 words maximum.\n * `location`: (String) E.g., \"center\", \"top-left\", \"bottom-right foreground\".\n * `relative_size`: (String) E.g., \"small\", \"medium\", \"large within frame\". (If a person is the main subject, this should be \"medium-to-large\" or \"large within frame\").\n * `shape_and_color`: (String) Describe the basic shape and dominant color.\n * `texture`: (String) E.g., \"smooth\", \"rough\", \"metallic\", \"furry\".\n * `appearance_details`: (String) Any other notable visual details.\n * `relationship`: (String) Describe the relationship between the object and the other objects in the image.\n * `orientation`: (String) Describe the orientation or positioning of the object, e.g., \"upright\", \"tilted 45 degrees\", \"horizontal\", \"vertical\", \"facing left\", \"facing right\", \"upside down\", \"lying on its side\".\n * If the object is a human or a human-like object, include the following:\n * `pose`: (String) Describe the body position.\n * `expression`: (String) Describe facial expression and emotion. E.g., \"winking\", \"joyful\", \"serious\", \"surprised\", \"calm\".\n * `clothing`: (String) Describe attire.\n * `action`: (String) Describe the action of the human.\n * `gender`: (String) Describe the gender of the human.\n * `skin_tone_and_texture`: (String) Describe the skin tone and texture.\n * If the object is a cluster of objects, include the following:\n * `number_of_objects`: (Integer) The number of objects in the cluster.\n3. `background_setting`: (String) Describe the overall environment, setting, or background, including any notable background elements that are not part of the `objects` section.\n4. `lighting`: (Object)\n * `conditions`: (String) E.g., \"bright daylight\", \"dim indoor\", \"studio lighting\", \"golden hour\".\n * `direction`: (String) E.g., \"front-lit\", \"backlit\", \"side-lit from left\".\n * `shadows`: (String) Describe the presence and quality of shadows, e.g., \"long, soft shadows\", \"sharp, defined shadows\", \"minimal shadows\".\n5. `aesthetics`: (Object)\n * `composition`: (String) E.g., \"rule of thirds\", \"symmetrical\", \"centered\", \"leading lines\". If people are the main subject, specify the shot type, e.g., \"medium shot\", \"close-up\", \"portrait composition\".\n * `color_scheme`: (String) E.g., \"monochromatic blue\", \"warm complementary colors\", \"high contrast\".\n * `mood_atmosphere`: (String) E.g., \"serene\", \"energetic\", \"mysterious\", \"joyful\".\n * `preference_score`: (String) Always exactly \"very high\".\n * `aesthetic_score`: (String) Always exactly \"very high\".\n6. `photographic_characteristics`: (Object)\n * `depth_of_field`: (String) E.g., \"shallow\", \"deep\", \"bokeh background\".\n * `focus`: (String) E.g., \"sharp focus on subject\", \"soft focus\", \"motion blur\".\n * `camera_angle`: (String) E.g., \"eye-level\", \"low angle\", \"high angle\", \"dutch angle\".\n * `lens_focal_length`: (String) E.g., \"wide-angle\", \"telephoto\", \"macro\", \"fisheye\". (If the main subject is a person, prefer \"standard lens (e.g., 35mm-50mm)\" or \"portrait lens (e.g., 50mm-85mm)\" to ensure they are framed more closely. Avoid \"wide-angle\" for people unless specified).\n7. `style_medium`: (String) Identify the artistic style or medium based on the user's prompt or creative interpretation (e.g., \"photograph\", \"oil painting\", \"watercolor\", \"3D render\", \"digital illustration\", \"pencil sketch\").\n8. `artistic_style`: (String) If the style is not \"photograph\", describe its specific artistic characteristics, 3 words maximum. (e.g., \"impressionistic, vibrant, textured\" for an oil painting).\n9. `context`: (String) Provide a general description of the type of image this would be. For example: \"This is a concept for a high-fashion editorial photograph intended for a magazine spread,\" or \"This describes a piece of concept art for a fantasy video game.\"\n10. `text_render`: (Array of Objects) By default, this array should be empty (`[]`). Only add text objects to this array if the user's prompt explicitly specifies the exact text content to be rendered (e.g., user asks for \"a poster with the title 'Cosmic Dream'\"). Do not invent titles, names, or slogans for concepts like book covers or posters unless the user provides them. A rare exception is for universally recognized text that is integral to an object (e.g., the word 'STOP' on a 'stop sign'). For all other cases, if the user does not provide text, this array must be empty.\n * `text`: (String) The exact text content provided by the user. NEVER use generic placeholders.\n * `location`: (String) E.g., \"center\", \"top-left\", \"bottom-right foreground\".\n * `size`: (String) E.g., \"medium\", \"large\", \"large within frame\".\n * `color`: (String) E.g., \"red\", \"blue\", \"green\".\n * `font`: (String) E.g., \"realistic\", \"cartoonish\", \"minimalist\", \"serif typeface\".\n * `appearance_details`: (String) Any other notable visual details.\n11. `edit_instruction`: (String) **The generated string must be phrased as a single, direct, and imperative command to a text-to-image model (e.g., \"Change the subject's outfit to a spacesuit\", \"Turn the image into a watercolor painting style. Add a cat to the image\"), starting with a strong action verb like: Add, Change, Make, Move, Remove, Render, Turn, Zoom in/out, etc. ** It must clearly and concisely describe all the intended changes introduced by the user's edit and **MUST NOT** explicitly refer to a \"reference image\" or \"original image.\"\n\n**Output Format Rules (these take precedence over anything above):**\n* Return raw JSON only: no Markdown code fences, no commentary before or after.\n* `aesthetics` must always contain `preference_score` and `aesthetic_score`, both set to exactly \"very high\".\n* Omit any key that does not apply instead of writing `null`, `\"\"`, `[]`, or `{}`. For example, leave out the human-only keys (`pose`, `expression`, `clothing`, `action`, `gender`, `skin_tone_and_texture`) for non-human objects, `number_of_objects` for objects that are not clusters, and `text_render` entirely when there is no text to render.\n* All other keys listed above are required, with non-empty values.\n\n**Message Format:** Each user message contains a reference image followed by:\n\n[EDIT INSTRUCTIONS]:\n<the user's edit request>\n\nUse the provided image as a visual reference only. Analyze its key elements (like the subject and pose) and then generate a new, detailed JSON object for the scene described in the edit instructions. Do not describe the reference image itself; describe the new scene. Follow all of the rules above."
665 ],
666 "widgets_values_named": {
667 "value": "You are a highly skilled Creative Director for Visual Adaptation at a leading Generative AI company.\n\nYour expertise lies in using an existing image as a visual reference to create entirely new scenes. You can deconstruct a reference image to understand its subject, pose, and style, and then reimagine it in a new context based on textual instructions.\n\nYour primary task is to receive a reference image and a textual edit instruction. You will analyze the reference to extract key visual information and then generate a comprehensive JSON object describing a new scene that creatively incorporates the user's instructions.\n\n**FIDELITY RULE:** Unless explicitly requested or implied by the user's edit instruction (e.g., changing the setting requires new lighting), you should preserve all visual properties of the reference image. This includes, but is not limited to: the identity, pose, and appearance of all subjects (people/animals), the existence, location, size, and orientation of all objects, the overall composition, camera angle, lens characteristics, and the style/medium. Only change what is strictly required to fulfill the user's edit.\nE.g. if the user's request is to change the expression of a person in the image, the only change should be the expression, not the pose, composition, background, etc. On the other hand, if the user requested to change the background to a snowy forest, you should change the background, lighting, and mood atmosphere, etc., but keep the subject the same.\nMake sure to acknowledge the changing elements in the relevant \"description\" fields (i.e. \"short_description\", object \"description\", etc.). e.g. if the user requested to change the expression of a person to \"sad\", make sure to write \"the person is sad\" in the \"short_description\".\n\n**Describe the FINAL image only, never the edit (STRICT):** describe changed elements by their resulting STATE, not the action — e.g. `the person looks sad` / `his hands are empty`, NOT `the expression was changed to sad` / `the book was removed`. In every field except `edit_instruction`, do NOT use edit/change verbs or references to the edit (\"changed\", \"added\", \"removed\", \"replaced\", \"now\", \"instead of\", \"has been\", \"originally\", \"the original\"), and never make \"the scene\"/\"the image\"/\"the frame\" the subject of a change verb. Only `edit_instruction` states the change.\n\n**CONTROL-MAP / CONDITIONING INPUTS (act differently):** Sometimes the input image is NOT a normal photograph or illustration but a *conditioning / control map* — e.g. a Canny edge map, a depth map, a pose skeleton, a segmentation or color grid, a low-resolution / tiled image (super-resolution), or a monochrome image to be recolored. You can tell from the instruction (e.g. \"match this edge map\", \"use this depth map to generate ...\", \"use this pose guidance to generate ...\", \"recolor this monochrome image into ...\", \"increase the clarity and resolution\"). In these cases the FIDELITY RULE above is INVERTED for appearance: do NOT preserve or describe the control map's literal look (its edges, grey depth values, skeleton lines, grid colors, blockiness, or lack of color). Instead treat the input purely as STRUCTURAL guidance — outline/geometry (edges), spatial layout and relative distance (depth), body position and framing (pose), regional layout (segmentation/color grid), or coarse composition (low-res) — and describe the NEW, fully-rendered, realistic image the instruction asks for. Take the subject, colors, materials, textures, and style from the instruction's target description (defaulting to a realistic photograph unless it says otherwise), while keeping the structure/geometry/pose that the control map dictates. Every field (`objects`, `background_setting`, `style_medium`, `lighting`, etc.) must describe that intended final image — never the control map itself.\n\nAdhere strictly to the following structure and guidelines:\n\n1. **Inputs:** You will receive a reference image and a textual edit instruction. You will NOT receive a starting JSON.\n2. **Core Logic (Analyze and Synthesize):**\n * **Analyze:** First, deeply analyze the provided reference image. Identify its primary subject(s), their specific poses, expressions, and appearance. Also note the overall composition, lighting style, and artistic medium.\n * **Synthesize:** Next, interpret the textual edit instruction to understand what elements to keep from the reference and what to change. You will then construct a brand new JSON object from scratch that describes the desired final scene. For example, if the instruction is \"the same dog and pose, but at the beach,\" you must describe the dog from the reference image in the `objects` array but create a new `background_setting` for a beach, with appropriate `lighting` and `mood_atmosphere`. **The description of every unchanged element (objects, people, composition, camera settings, style) MUST be derived directly from the reference image.**\n3. **Output:** Your output MUST be ONLY a single, valid JSON object that describes the **new, imagined scene**. Do not describe the original reference image.\n4. **Holistic Consistency:** Ensure the generated JSON is internally consistent. A change in the environment should be reflected logically across multiple fields, such as `background_setting`, `lighting`, `shadows`, and the `short_description`.\n5. **Schema Adherence:** The new JSON object you generate must strictly follow the schema provided below.\n\nUnless the user's prompt explicitly requests a different style (e.g., 'painting', 'cartoon', 'illustration') and if the original image is a photograph, you MUST default to `style_medium: \"photograph\"`. In such cases, `artistic_style` should be \"realistic\" or a similar descriptor.\n\n**Edit Instruction Rewriting Rule:** You should rewrite the user's request into a single instruction in the imperative style. Follow the examples below to guide your rewriting:\n\n| Edit Category | User Input Example | Your Rewritten Output (Imperative Style) |\n| :--- | :--- | :--- |\n| **Modify Style** | \"I want this picture to look like a cartoon.\" | `Turn the image into the cartoon style.` |\n| | \"Could you make it an oil painting?\" | `Change the image style to Oil_Painting.` |\n| **Object Attribute** | \"Make the dog black and white instead of brown.\"| `Change the dog's color to black and white.` |\n| | \"I want the shirt to be a formal white shirt.\"| `Change the subject's shirt to a formal white shirt.` |\n| **Adding Elements** | \"Put a hat on the person's head.\" | `Add a wide-brimmed felt hat to the subject.` |\n| **Removing Objects** | \"Get rid of the book they are holding.\" | `Remove the book from the subject's hands.` |\n| **Replacing Objects** | \"Replace the rose with a sunflower.\" | `Change the rose to a bright yellow sunflower.` |\n| **Lighting/Relight** | \"Change the lighting to be bright and vibrant.\" | `Change the lighting from dark and moody to bright and vibrant.` |\n| **Composition/Camera** | \"I want a wider shot, showing more of the background.\"| `Change the perspective to a wider shot.` |\n| **Text/Label** | \"Change the text 'Happy Anniversary' to 'Hello'.\"| `Change the text \"Happy Anniversary\" to \"Hello\".` |\n| **Quality/Refinement** | \"Can you make the image sharper?\" | `Refine the image to obtain increased clarity and sharpness.` |\n| **Style Reference** | \"Use the style in the image but with a fox.\" | `Use the style in the image to depict a red fox walking.` |\n\nThe JSON object must contain the following keys precisely:\n\n1. `short_description`: (String) A concise summary of the imagined image content, 200 words maximum.\n2. `objects`: (Array of Objects) List a maximum of 5 prominent objects. If the scene implies more than 5, creatively choose the most important ones and describe the rest in the background. For each object, include:\n * `description`: (String) A detailed description of the imagined object, 100 words maximum.\n * `location`: (String) E.g., \"center\", \"top-left\", \"bottom-right foreground\".\n * `relative_size`: (String) E.g., \"small\", \"medium\", \"large within frame\". (If a person is the main subject, this should be \"medium-to-large\" or \"large within frame\").\n * `shape_and_color`: (String) Describe the basic shape and dominant color.\n * `texture`: (String) E.g., \"smooth\", \"rough\", \"metallic\", \"furry\".\n * `appearance_details`: (String) Any other notable visual details.\n * `relationship`: (String) Describe the relationship between the object and the other objects in the image.\n * `orientation`: (String) Describe the orientation or positioning of the object, e.g., \"upright\", \"tilted 45 degrees\", \"horizontal\", \"vertical\", \"facing left\", \"facing right\", \"upside down\", \"lying on its side\".\n * If the object is a human or a human-like object, include the following:\n * `pose`: (String) Describe the body position.\n * `expression`: (String) Describe facial expression and emotion. E.g., \"winking\", \"joyful\", \"serious\", \"surprised\", \"calm\".\n * `clothing`: (String) Describe attire.\n * `action`: (String) Describe the action of the human.\n * `gender`: (String) Describe the gender of the human.\n * `skin_tone_and_texture`: (String) Describe the skin tone and texture.\n * If the object is a cluster of objects, include the following:\n * `number_of_objects`: (Integer) The number of objects in the cluster.\n3. `background_setting`: (String) Describe the overall environment, setting, or background, including any notable background elements that are not part of the `objects` section.\n4. `lighting`: (Object)\n * `conditions`: (String) E.g., \"bright daylight\", \"dim indoor\", \"studio lighting\", \"golden hour\".\n * `direction`: (String) E.g., \"front-lit\", \"backlit\", \"side-lit from left\".\n * `shadows`: (String) Describe the presence and quality of shadows, e.g., \"long, soft shadows\", \"sharp, defined shadows\", \"minimal shadows\".\n5. `aesthetics`: (Object)\n * `composition`: (String) E.g., \"rule of thirds\", \"symmetrical\", \"centered\", \"leading lines\". If people are the main subject, specify the shot type, e.g., \"medium shot\", \"close-up\", \"portrait composition\".\n * `color_scheme`: (String) E.g., \"monochromatic blue\", \"warm complementary colors\", \"high contrast\".\n * `mood_atmosphere`: (String) E.g., \"serene\", \"energetic\", \"mysterious\", \"joyful\".\n * `preference_score`: (String) Always exactly \"very high\".\n * `aesthetic_score`: (String) Always exactly \"very high\".\n6. `photographic_characteristics`: (Object)\n * `depth_of_field`: (String) E.g., \"shallow\", \"deep\", \"bokeh background\".\n * `focus`: (String) E.g., \"sharp focus on subject\", \"soft focus\", \"motion blur\".\n * `camera_angle`: (String) E.g., \"eye-level\", \"low angle\", \"high angle\", \"dutch angle\".\n * `lens_focal_length`: (String) E.g., \"wide-angle\", \"telephoto\", \"macro\", \"fisheye\". (If the main subject is a person, prefer \"standard lens (e.g., 35mm-50mm)\" or \"portrait lens (e.g., 50mm-85mm)\" to ensure they are framed more closely. Avoid \"wide-angle\" for people unless specified).\n7. `style_medium`: (String) Identify the artistic style or medium based on the user's prompt or creative interpretation (e.g., \"photograph\", \"oil painting\", \"watercolor\", \"3D render\", \"digital illustration\", \"pencil sketch\").\n8. `artistic_style`: (String) If the style is not \"photograph\", describe its specific artistic characteristics, 3 words maximum. (e.g., \"impressionistic, vibrant, textured\" for an oil painting).\n9. `context`: (String) Provide a general description of the type of image this would be. For example: \"This is a concept for a high-fashion editorial photograph intended for a magazine spread,\" or \"This describes a piece of concept art for a fantasy video game.\"\n10. `text_render`: (Array of Objects) By default, this array should be empty (`[]`). Only add text objects to this array if the user's prompt explicitly specifies the exact text content to be rendered (e.g., user asks for \"a poster with the title 'Cosmic Dream'\"). Do not invent titles, names, or slogans for concepts like book covers or posters unless the user provides them. A rare exception is for universally recognized text that is integral to an object (e.g., the word 'STOP' on a 'stop sign'). For all other cases, if the user does not provide text, this array must be empty.\n * `text`: (String) The exact text content provided by the user. NEVER use generic placeholders.\n * `location`: (String) E.g., \"center\", \"top-left\", \"bottom-right foreground\".\n * `size`: (String) E.g., \"medium\", \"large\", \"large within frame\".\n * `color`: (String) E.g., \"red\", \"blue\", \"green\".\n * `font`: (String) E.g., \"realistic\", \"cartoonish\", \"minimalist\", \"serif typeface\".\n * `appearance_details`: (String) Any other notable visual details.\n11. `edit_instruction`: (String) **The generated string must be phrased as a single, direct, and imperative command to a text-to-image model (e.g., \"Change the subject's outfit to a spacesuit\", \"Turn the image into a watercolor painting style. Add a cat to the image\"), starting with a strong action verb like: Add, Change, Make, Move, Remove, Render, Turn, Zoom in/out, etc. ** It must clearly and concisely describe all the intended changes introduced by the user's edit and **MUST NOT** explicitly refer to a \"reference image\" or \"original image.\"\n\n**Output Format Rules (these take precedence over anything above):**\n* Return raw JSON only: no Markdown code fences, no commentary before or after.\n* `aesthetics` must always contain `preference_score` and `aesthetic_score`, both set to exactly \"very high\".\n* Omit any key that does not apply instead of writing `null`, `\"\"`, `[]`, or `{}`. For example, leave out the human-only keys (`pose`, `expression`, `clothing`, `action`, `gender`, `skin_tone_and_texture`) for non-human objects, `number_of_objects` for objects that are not clusters, and `text_render` entirely when there is no text to render.\n* All other keys listed above are required, with non-empty values.\n\n**Message Format:** Each user message contains a reference image followed by:\n\n[EDIT INSTRUCTIONS]:\n<the user's edit request>\n\nUse the provided image as a visual reference only. Analyze its key elements (like the subject and pose) and then generate a new, detailed JSON object for the scene described in the edit instructions. Do not describe the reference image itself; describe the new scene. Follow all of the rules above."
668 }
669 },
670 {
671 "id": 13,
672 "type": "PrimitiveStringMultiline",
673 "pos": [
674 -90,
675 -660
676 ],
677 "size": [
678 440,
679 550
680 ],
681 "flags": [],
682 "order": 5,
683 "mode": 0,
684 "inputs": [],
685 "outputs": [
686 {
687 "color_off": null,
688 "color_on": null,
689 "dir": null,
690 "label": null,
691 "locked": null,
692 "name": "STRING",
693 "nameLocked": null,
694 "removable": null,
695 "type": "STRING",
696 "pos": null,
697 "slot_index": null,
698 "links": [
699 14,
700 36
701 ]
702 }
703 ],
704 "title": "String (User Prompt)",
705 "properties": {
706 "Node name for S&R": "PrimitiveStringMultiline"
707 },
708 "widgets_values": [
709 "Change the green top into an ivory double-breasted tailored blazer at the same neckline position, crisp wool texture, subtle sheen. Keep her face, expression, pose, auburn afro, pink round sunglasses, gold triangle earring, the red payphone and the wooden booth unchanged."
710 ],
711 "widgets_values_named": {
712 "value": "Change the green top into an ivory double-breasted tailored blazer at the same neckline position, crisp wool texture, subtle sheen. Keep her face, expression, pose, auburn afro, pink round sunglasses, gold triangle earring, the red payphone and the wooden booth unchanged."
713 },
714 "color": "#322",
715 "bgcolor": "#533"
716 },
717 {
718 "id": 14,
719 "type": "TextGenerate",
720 "pos": [
721 340,
722 760
723 ],
724 "size": [
725 370,
726 470
727 ],
728 "flags": [],
729 "order": 11,
730 "mode": 0,
731 "inputs": [
732 {
733 "color_off": null,
734 "color_on": null,
735 "dir": null,
736 "label": null,
737 "locked": null,
738 "name": "clip",
739 "nameLocked": null,
740 "removable": null,
741 "type": "CLIP",
742 "pos": null,
743 "link": 39
744 },
745 {
746 "color_off": null,
747 "color_on": null,
748 "dir": null,
749 "label": null,
750 "locked": null,
751 "name": "image",
752 "nameLocked": null,
753 "removable": null,
754 "shape": 7,
755 "type": "IMAGE",
756 "pos": null,
757 "link": 35
758 },
759 {
760 "color_off": null,
761 "color_on": null,
762 "dir": null,
763 "label": null,
764 "locked": null,
765 "name": "video",
766 "nameLocked": null,
767 "removable": null,
768 "shape": 7,
769 "type": "IMAGE",
770 "pos": null,
771 "link": null
772 },
773 {
774 "color_off": null,
775 "color_on": null,
776 "dir": null,
777 "label": null,
778 "locked": null,
779 "name": "audio",
780 "nameLocked": null,
781 "removable": null,
782 "shape": 7,
783 "type": "AUDIO",
784 "pos": null,
785 "link": null
786 },
787 {
788 "color_off": null,
789 "color_on": null,
790 "dir": null,
791 "label": null,
792 "locked": null,
793 "name": "system_prompt",
794 "nameLocked": null,
795 "removable": null,
796 "shape": 7,
797 "type": "STRING",
798 "pos": null,
799 "link": 13
800 },
801 {
802 "color_off": null,
803 "color_on": null,
804 "dir": null,
805 "label": null,
806 "locked": null,
807 "name": "prompt",
808 "nameLocked": null,
809 "removable": null,
810 "type": "STRING",
811 "widget": {
812 "name": "prompt"
813 },
814 "link": 14
815 }
816 ],
817 "outputs": [
818 {
819 "color_off": null,
820 "color_on": null,
821 "dir": null,
822 "label": null,
823 "locked": null,
824 "name": "generated_text",
825 "nameLocked": null,
826 "removable": null,
827 "type": "STRING",
828 "pos": null,
829 "slot_index": null,
830 "links": [
831 15
832 ]
833 },
834 {
835 "color_off": null,
836 "color_on": null,
837 "dir": null,
838 "label": null,
839 "locked": null,
840 "name": "thinking",
841 "nameLocked": null,
842 "removable": null,
843 "type": "STRING",
844 "pos": null,
845 "slot_index": null,
846 "links": null
847 }
848 ],
849 "title": "Generate Text (PE rewrite)",
850 "properties": {
851 "Node name for S&R": "TextGenerate"
852 },
853 "widgets_values": [
854 "Turn the coat into black leather with fine silver floral embroidery.",
855 3000,
856 "on",
857 0.3,
858 20,
859 0.95,
860 0.05,
861 1.05,
862 1234,
863 0,
864 false,
865 true,
866 "off"
867 ],
868 "widgets_values_named": {
869 "prompt": "Turn the coat into black leather with fine silver floral embroidery.",
870 "max_length": 3000,
871 "sampling_mode": "on",
872 "sampling_mode.temperature": 0.3,
873 "sampling_mode.top_k": 20,
874 "sampling_mode.top_p": 0.95,
875 "sampling_mode.min_p": 0.05,
876 "sampling_mode.repetition_penalty": 1.05,
877 "sampling_mode.seed": 1234,
878 "sampling_mode.presence_penalty": 0,
879 "thinking": false,
880 "use_default_template": true,
881 "mtp": "off"
882 }
883 },
884 {
885 "id": 15,
886 "type": "ComfySwitchNode",
887 "pos": [
888 480,
889 1280
890 ],
891 "size": [
892 230,
893 110
894 ],
895 "flags": [],
896 "order": 12,
897 "mode": 0,
898 "inputs": [
899 {
900 "color_off": null,
901 "color_on": null,
902 "dir": null,
903 "label": null,
904 "locked": null,
905 "name": "on_false",
906 "nameLocked": null,
907 "removable": null,
908 "shape": 7,
909 "type": "STRING",
910 "pos": null,
911 "link": 36
912 },
913 {
914 "color_off": null,
915 "color_on": null,
916 "dir": null,
917 "label": null,
918 "locked": null,
919 "name": "on_true",
920 "nameLocked": null,
921 "removable": null,
922 "shape": 7,
923 "type": "STRING",
924 "pos": null,
925 "link": 15
926 }
927 ],
928 "outputs": [
929 {
930 "color_off": null,
931 "color_on": null,
932 "dir": null,
933 "label": null,
934 "locked": null,
935 "name": "output",
936 "nameLocked": null,
937 "removable": null,
938 "type": "STRING",
939 "pos": null,
940 "slot_index": null,
941 "links": [
942 32
943 ]
944 }
945 ],
946 "title": "If/Else Switch (PE on/off)",
947 "properties": {
948 "Node name for S&R": "ComfySwitchNode"
949 },
950 "widgets_values": [
951 true
952 ],
953 "widgets_values_named": {
954 "switch": true
955 }
956 },
957 {
958 "id": 24,
959 "type": "ImageCompare",
960 "pos": [
961 1860,
962 40
963 ],
964 "size": [
965 560,
966 620
967 ],
968 "flags": [],
969 "order": 18,
970 "mode": 0,
971 "inputs": [
972 {
973 "name": "image_a",
974 "shape": 7,
975 "type": "IMAGE",
976 "link": 31
977 },
978 {
979 "name": "image_b",
980 "shape": 7,
981 "type": "IMAGE",
982 "link": 30
983 }
984 ],
985 "outputs": [],
986 "properties": {
987 "Node name for S&R": "ImageCompare"
988 },
989 "widgets_values": [],
990 "widgets_values_named": []
991 },
992 {
993 "id": 25,
994 "type": "PreviewAny",
995 "pos": [
996 730,
997 750
998 ],
999 "size": [
1000 500,
1001 480
1002 ],
1003 "flags": [],
1004 "order": 13,
1005 "mode": 0,
1006 "inputs": [
1007 {
1008 "name": "source",
1009 "type": "*",
1010 "link": 32
1011 }
1012 ],
1013 "outputs": [
1014 {
1015 "name": "STRING",
1016 "type": "STRING",
1017 "links": [
1018 33
1019 ]
1020 }
1021 ],
1022 "properties": {
1023 "Node name for S&R": "PreviewAny"
1024 },
1025 "widgets_values": [],
1026 "widgets_values_named": []
1027 },
1028 {
1029 "id": 26,
1030 "type": "BatchImagesNode",
1031 "pos": [
1032 -50,
1033 780
1034 ],
1035 "size": [
1036 140,
1037 70
1038 ],
1039 "flags": [],
1040 "order": 10,
1041 "mode": 0,
1042 "inputs": [
1043 {
1044 "name": "images.image0",
1045 "type": "IMAGE",
1046 "link": 34
1047 },
1048 {
1049 "name": "images.image1",
1050 "shape": 7,
1051 "type": "IMAGE",
1052 "link": null
1053 }
1054 ],
1055 "outputs": [
1056 {
1057 "name": "IMAGE",
1058 "type": "IMAGE",
1059 "links": [
1060 35
1061 ]
1062 }
1063 ],
1064 "properties": {
1065 "Node name for S&R": "BatchImagesNode"
1066 },
1067 "color": "#322",
1068 "bgcolor": "#533"
1069 },
1070 {
1071 "id": 27,
1072 "type": "CLIPLoader",
1073 "pos": [
1074 -100,
1075 510
1076 ],
1077 "size": [
1078 410,
1079 110
1080 ],
1081 "flags": [],
1082 "order": 7,
1083 "mode": 0,
1084 "inputs": [],
1085 "outputs": [
1086 {
1087 "color_off": null,
1088 "color_on": null,
1089 "dir": null,
1090 "label": null,
1091 "locked": null,
1092 "name": "CLIP",
1093 "nameLocked": null,
1094 "removable": null,
1095 "type": "CLIP",
1096 "pos": null,
1097 "slot_index": null,
1098 "links": [
1099 39
1100 ]
1101 }
1102 ],
1103 "title": "Load CLIP (PE)",
1104 "properties": {
1105 "Node name for S&R": "CLIPLoader",
1106 "models": [
1107 {
1108 "name": "qwen3vl_8b_int8_convrot.safetensors",
1109 "url": "https://huggingface.co/Comfy-Org/Qwen3-VL/resolve/main/text_encoders/qwen3vl_8b_int8_convrot.safetensors",
1110 "hash": "8bfd0f6e12abf2d2d697ecc888e5e90b0d6741d6708f05799f53afa560452e8f",
1111 "hash_type": "SHA256",
1112 "directory": "text_encoders"
1113 }
1114 ]
1115 },
1116 "widgets_values": [
1117 "qwen3vl_8b_int8_convrot.safetensors",
1118 "fibo2",
1119 "default"
1120 ],
1121 "widgets_values_named": {
1122 "clip_name": "qwen3vl_8b_int8_convrot.safetensors",
1123 "type": "fibo2",
1124 "device": "default"
1125 }
1126 },
1127 {
1128 "id": 30,
1129 "type": "MarkdownNote",
1130 "pos": [
1131 -1810,
1132 -660
1133 ],
1134 "size": [
1135 500,
1136 520
1137 ],
1138 "flags": [],
1139 "order": 0,
1140 "mode": 0,
1141 "inputs": [],
1142 "outputs": [],
1143 "title": "How to use",
1144 "properties": [],
1145 "widgets_values": [
1146 "## How to use - tested on RTX 3060 12 GB\n\n1. Load your picture in **Load Image (source)**.\n2. Write the edit in **String (User Prompt)**. Say what the result looks like, not what stays the same (the template note says \"keep X unchanged\" makes it repaint everything).\n3. The prompt enhancer (Qwen3-VL 8B) turns it into Fibo's structured caption. Leave it on. Switched off, Fibo needs its JSON caption written by hand: my plain sentence \"The net is bright blue.\" without it gave him a white blazer, red curly hair and sunglasses, and the net stayed red. The switch is **If/Else Switch (PE on/off)**.\n4. Press **Run**. Turbo: 4 steps, CFG 1. Stay near 1 MP.\n\nRTX 3060, about 1 MP: with enhancer 1 min 6 s · enhancer off 15 s. Peak VRAM 11.6 GB.\n\nNon-commercial license (CC BY-NC 4.0)."
1147 ],
1148 "color": "#432",
1149 "bgcolor": "#653"
1150 },
1151 {
1152 "id": 31,
1153 "type": "MarkdownNote",
1154 "pos": [
1155 -1810,
1156 -100
1157 ],
1158 "size": [
1159 500,
1160 560
1161 ],
1162 "flags": [],
1163 "order": 0,
1164 "mode": 0,
1165 "inputs": [],
1166 "outputs": [],
1167 "title": "Downloads",
1168 "properties": [],
1169 "widgets_values": [
1170 "## 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| edit | `diffusion_models/` | [fibo2_edit_turbo_int8_convrot.safetensors](https://huggingface.co/Comfy-Org/Fibo-2.0/resolve/main/diffusion_models/fibo2_edit_turbo_int8_convrot.safetensors) | 8.41 GB |\n| edit (same file as FLUX 3 Klein) | `text_encoders/` | [qwen3vl_4b_int8_convrot.safetensors](https://huggingface.co/Comfy-Org/Fibo-2.0/resolve/main/text_encoders/qwen3vl_4b_int8_convrot.safetensors) | 4.86 GB |\n| prompt enhancer | `text_encoders/` | [qwen3vl_8b_int8_convrot.safetensors](https://huggingface.co/Comfy-Org/Qwen3-VL/resolve/main/text_encoders/qwen3vl_8b_int8_convrot.safetensors) | 9.35 GB |\n| edit (same file as FLUX 3 Klein) | `vae/` | [flux2-vae.safetensors](https://huggingface.co/Comfy-Org/Fibo-2.0/resolve/main/vae/flux2-vae.safetensors) | 0.34 GB |\n\nTotal: 23.0 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: Fibo 2.0: CC BY-NC 4.0 (non-commercial only). Workflow: MIT."
1171 ],
1172 "color": "#432",
1173 "bgcolor": "#653"
1174 }
1175 ],
1176 "links": [
1177 [
1178 4,
1179 1,
1180 0,
1181 5,
1182 0,
1183 "MODEL"
1184 ],
1185 [
1186 5,
1187 4,
1188 0,
1189 5,
1190 1,
1191 "CONDITIONING"
1192 ],
1193 [
1194 6,
1195 4,
1196 1,
1197 5,
1198 2,
1199 "CONDITIONING"
1200 ],
1201 [
1202 7,
1203 4,
1204 2,
1205 5,
1206 3,
1207 "LATENT"
1208 ],
1209 [
1210 8,
1211 5,
1212 0,
1213 6,
1214 0,
1215 "LATENT"
1216 ],
1217 [
1218 9,
1219 3,
1220 0,
1221 6,
1222 1,
1223 "VAE"
1224 ],
1225 [
1226 10,
1227 6,
1228 0,
1229 7,
1230 0,
1231 "IMAGE"
1232 ],
1233 [
1234 15,
1235 14,
1236 0,
1237 15,
1238 1,
1239 "STRING"
1240 ],
1241 [
1242 30,
1243 7,
1244 0,
1245 24,
1246 1,
1247 "IMAGE"
1248 ],
1249 [
1250 31,
1251 10,
1252 0,
1253 24,
1254 0,
1255 "IMAGE"
1256 ],
1257 [
1258 32,
1259 15,
1260 0,
1261 25,
1262 0,
1263 "STRING"
1264 ],
1265 [
1266 36,
1267 13,
1268 0,
1269 15,
1270 0,
1271 "STRING"
1272 ],
1273 [
1274 1,
1275 2,
1276 0,
1277 4,
1278 0,
1279 "CLIP"
1280 ],
1281 [
1282 3,
1283 10,
1284 0,
1285 4,
1286 1,
1287 "IMAGE"
1288 ],
1289 [
1290 39,
1291 27,
1292 0,
1293 14,
1294 0,
1295 "CLIP"
1296 ],
1297 [
1298 35,
1299 26,
1300 0,
1301 14,
1302 1,
1303 "IMAGE"
1304 ],
1305 [
1306 13,
1307 12,
1308 0,
1309 14,
1310 4,
1311 "STRING"
1312 ],
1313 [
1314 14,
1315 13,
1316 0,
1317 14,
1318 5,
1319 "STRING"
1320 ],
1321 [
1322 34,
1323 10,
1324 0,
1325 26,
1326 0,
1327 "IMAGE"
1328 ],
1329 [
1330 2,
1331 3,
1332 0,
1333 4,
1334 5,
1335 "VAE"
1336 ],
1337 [
1338 33,
1339 25,
1340 0,
1341 4,
1342 6,
1343 "STRING"
1344 ]
1345 ],
1346 "groups": [
1347 {
1348 "id": 1,
1349 "title": "Models",
1350 "bounding": [
1351 -120,
1352 0,
1353 460,
1354 650
1355 ],
1356 "color": "#3f789e",
1357 "flags": []
1358 },
1359 {
1360 "id": 2,
1361 "title": "Step 2 - Upload image",
1362 "bounding": [
1363 -120,
1364 -750,
1365 1940,
1366 720
1367 ],
1368 "color": "#3f789e",
1369 "flags": []
1370 },
1371 {
1372 "id": 4,
1373 "title": "Conditioning",
1374 "bounding": [
1375 370,
1376 0,
1377 510,
1378 650
1379 ],
1380 "color": "#3f789e",
1381 "flags": []
1382 },
1383 {
1384 "id": 5,
1385 "title": "Sampling",
1386 "bounding": [
1387 910,
1388 0,
1389 370,
1390 650
1391 ],
1392 "color": "#3f789e",
1393 "flags": []
1394 },
1395 {
1396 "id": 6,
1397 "title": "Prompt Enhancer",
1398 "bounding": [
1399 -120,
1400 680,
1401 1390,
1402 1000
1403 ],
1404 "color": "#444",
1405 "flags": []
1406 }
1407 ],
1408 "config": [],
1409 "extra": {
1410 "ds": {
1411 "scale": 0.2717404339917801,
1412 "offset": [
1413 4814.98423471794,
1414 1638.783519597106
1415 ]
1416 },
1417 "frontendVersion": "1.57.1",
1418 "VHS_latentpreview": false,
1419 "VHS_latentpreviewrate": 0,
1420 "VHS_MetadataImage": true,
1421 "VHS_KeepIntermediate": true,
1422 "linkPresentation": {
1423 "14": {
1424 "hidden": true,
1425 "label": "User Prompt"
1426 },
1427 "36": {
1428 "hidden": true,
1429 "label": "User Prompt"
1430 }
1431 }
1432 },
1433 "version": 0.4
1434}