ComfyPilotComfyPilot
工作流 / 图像 / Seedream 5.5：图像编辑
Seedream 5.5：图像编辑

Seedream 5.5：图像编辑

Seedream 5.5: Image Edit

上传一张图像和一段提示词即可编辑。输出保持原图分辨率：工作流从输入图像读取尺寸，因此编辑后的图像与原图宽高一致。

图像 7 个节点 1 个输入 980 次使用 ComfyUI ≥ 0.45.0 ComfyUI 10 KB

下载工作流 JSON 查看官方来源

下载后把文件拖进 ComfyUI 窗口即可打开。

节点构成

MarkdownNote × 2
LoadImage × 1
SaveImage × 1
ImageCompare × 1
ByteDanceSeedreamNodeV3 × 1
GetImageSize × 1

官方标注的模型

Seedream 5.5

工作流 JSON

和上面「下载工作流 JSON」是同一份文件。

1{
2 "id": "f6e5e4bd-0098-4d73-92c4-c5b015bfc5c2",
3 "revision": 0,
4 "last_node_id": 42,
5 "last_link_id": 24,
6 "nodes": [
7 {
8 "id": 11,
9 "type": "LoadImage",
10 "pos": [
11 100,
12 130
13 ],
14 "size": [
15 465.86462125786784,
16 603.3658198813556
17 ],
18 "flags": [],
19 "order": 0,
20 "mode": 0,
21 "inputs": [],
22 "outputs": [
23 {
24 "name": "IMAGE",
25 "type": "IMAGE",
26 "links": [
27 18,
28 20,
29 22
30 ]
31 },
32 {
33 "name": "MASK",
34 "type": "MASK",
35 "links": null
36 }
37 ],
38 "properties": {
39 "Node name for S&R": "LoadImage"
40 },
41 "widgets_values": [
42 "api_bytedance_seedream_5_5_image_edit_input.png",
43 "image"
44 ],
45 "widgets_values_named": {
46 "image": "api_bytedance_seedream_5_5_image_edit_input.png",
47 "upload": "image"
48 }
49 },
50 {
51 "id": 32,
52 "type": "SaveImage",
53 "pos": [
54 1061.8646212578678,
55 130
56 ],
57 "size": [
58 520.7864005025971,
59 622.2658758022359
60 ],
61 "flags": [],
62 "order": 5,
63 "mode": 0,
64 "inputs": [
65 {
66 "name": "images",
67 "type": "IMAGE",
68 "link": 21
69 }
70 ],
71 "outputs": [
72 {
73 "name": "images",
74 "type": "IMAGE",
75 "links": [
76 19
77 ]
78 }
79 ],
80 "properties": [],
81 "widgets_values": [
82 "api_bytedance_seedream_5_5_image_edit"
83 ],
84 "widgets_values_named": {
85 "filename_prefix": "api_bytedance_seedream_5_5_image_edit"
86 }
87 },
88 {
89 "id": 38,
90 "type": "ImageCompare",
91 "pos": [
92 1628.5593409220894,
93 140.59820813792882
94 ],
95 "size": [
96 584.0471428072578,
97 604.0250980112673
98 ],
99 "flags": [],
100 "order": 6,
101 "mode": 0,
102 "inputs": [
103 {
104 "name": "image_a",
105 "shape": 7,
106 "type": "IMAGE",
107 "link": 18
108 },
109 {
110 "name": "image_b",
111 "shape": 7,
112 "type": "IMAGE",
113 "link": 19
114 }
115 ],
116 "outputs": [],
117 "properties": {
118 "Node name for S&R": "ImageCompare"
119 },
120 "widgets_values": [],
121 "widgets_values_named": []
122 },
123 {
124 "id": 39,
125 "type": "MarkdownNote",
126 "pos": [
127 -1100,
128 130
129 ],
130 "size": [
131 560,
132 520
133 ],
134 "flags": [],
135 "order": 1,
136 "mode": 0,
137 "inputs": [],
138 "outputs": [],
139 "properties": [],
140 "widgets_values": [
141 "## Seedream 5.5 Image Edit\n\n**Input Assets**\n- `api_bytedance_seedream_5_5_image_edit_input.png` — source portrait\n\nEdit a picture with Seedream 5.5 through the partner API: state what changes, repeat what must stay. The node accepts two reference images; the first one is wired here.\n\n1. **Load image** — the source picture.\n2. **Edit** — the prompt names the changes (chocolate → macaron, hair, earrings, top, background) and restates the invariants (face, smile, pose, hand, framing).\n3. **Compare / Save** — the result sits next to the input.\n\n**Size** — the output keeps the source resolution: `Get Image Size` feeds the node's width and height, so a 1024×1024 picture comes back at 1024×1024. See the size note beside the canvas for preset sizes.\n"
142 ],
143 "widgets_values_named": {
144 "text": "## Seedream 5.5 Image Edit\n\n**Input Assets**\n- `api_bytedance_seedream_5_5_image_edit_input.png` — source portrait\n\nEdit a picture with Seedream 5.5 through the partner API: state what changes, repeat what must stay. The node accepts two reference images; the first one is wired here.\n\n1. **Load image** — the source picture.\n2. **Edit** — the prompt names the changes (chocolate → macaron, hair, earrings, top, background) and restates the invariants (face, smile, pose, hand, framing).\n3. **Compare / Save** — the result sits next to the input.\n\n**Size** — the output keeps the source resolution: `Get Image Size` feeds the node's width and height, so a 1024×1024 picture comes back at 1024×1024. See the size note beside the canvas for preset sizes.\n"
145 },
146 "color": "#222",
147 "bgcolor": "#000"
148 },
149 {
150 "id": 40,
151 "type": "ByteDanceSeedreamNodeV3",
152 "pos": [
153 610.7732142282116,
154 132.24348323738076
155 ],
156 "size": [
157 400,
158 324
159 ],
160 "flags": [],
161 "order": 4,
162 "mode": 0,
163 "inputs": [
164 {
165 "name": "model.width",
166 "type": "INT",
167 "widget": {
168 "name": "model.width"
169 },
170 "link": 23
171 },
172 {
173 "name": "model.height",
174 "type": "INT",
175 "widget": {
176 "name": "model.height"
177 },
178 "link": 24
179 },
180 {
181 "name": "model.images.image_1",
182 "shape": 7,
183 "type": "IMAGE",
184 "link": 20
185 },
186 {
187 "name": "model.images.image_2",
188 "shape": 7,
189 "type": "IMAGE",
190 "link": null
191 }
192 ],
193 "outputs": [
194 {
195 "name": "IMAGE",
196 "type": "IMAGE",
197 "links": [
198 21
199 ]
200 }
201 ],
202 "properties": {
203 "Node name for S&R": "ByteDanceSeedreamNodeV3"
204 },
205 "widgets_values": [
206 "Change the chocolate to a lilac macaron, hair to a chin-length bob, earrings to pearl studs, top to lavender ruffle blouse, background to soft lilac. Keep the same face, smile, pose, hand covering the left eye, and framing.\n\nKeep same face, smile, pose, and framing. No text.",
207 "seedream 5.5",
208 "Custom",
209 2048,
210 2048,
211 "auto",
212 42,
213 "randomize",
214 false,
215 "randomize"
216 ],
217 "widgets_values_named": {
218 "prompt": "Change the chocolate to a lilac macaron, hair to a chin-length bob, earrings to pearl studs, top to lavender ruffle blouse, background to soft lilac. Keep the same face, smile, pose, hand covering the left eye, and framing.\n\nKeep same face, smile, pose, and framing. No text.",
219 "model": "seedream 5.5",
220 "model.size_preset": "Custom",
221 "model.width": 2048,
222 "model.height": 2048,
223 "model.background": "auto",
224 "model.seed": 42,
225 "control_after_generate#1": "randomize",
226 "model.watermark": false,
227 "control_after_generate": "randomize"
228 },
229 "color": "#432",
230 "bgcolor": "#653"
231 },
232 {
233 "id": 41,
234 "type": "GetImageSize",
235 "pos": [
236 620.179573544103,
237 519.3658003143698
238 ],
239 "size": [
240 140,
241 66
242 ],
243 "flags": [],
244 "order": 3,
245 "mode": 0,
246 "inputs": [
247 {
248 "name": "image",
249 "type": "IMAGE",
250 "link": 22
251 }
252 ],
253 "outputs": [
254 {
255 "name": "width",
256 "type": "INT",
257 "links": [
258 23
259 ]
260 },
261 {
262 "name": "height",
263 "type": "INT",
264 "links": [
265 24
266 ]
267 },
268 {
269 "name": "batch_size",
270 "type": "INT",
271 "links": null
272 }
273 ],
274 "properties": {
275 "Node name for S&R": "GetImageSize"
276 }
277 },
278 {
279 "id": 42,
280 "type": "MarkdownNote",
281 "pos": [
282 -500,
283 130
284 ],
285 "size": [
286 520,
287 440
288 ],
289 "flags": [],
290 "order": 2,
291 "mode": 0,
292 "inputs": [],
293 "outputs": [],
294 "title": "Note: Size",
295 "properties": [],
296 "widgets_values": [
297 "**Size — custom, taken from the input image**\n\n`Get Image Size` reads the loaded picture and feeds `Size preset = Custom` plus that exact width and height, so the edit keeps the source resolution and aspect ratio.\n\nTo use a recommended size instead:\n1. Delete the `Get Image Size` node (its width/height links disappear with it).\n2. Pick a preset in `Size preset`, e.g. `(1K) 1024x1024 (1:1)`, `(1.5K) 1328x1776 (3:4)`, `(2K) 2048x2048 (1:1)`.\n3. Or keep `Custom` and type `width` / `height` yourself.\n\nA preset re-composes the canvas at that size; the edit still follows the source picture.\n"
298 ],
299 "widgets_values_named": {
300 "text": "**Size — custom, taken from the input image**\n\n`Get Image Size` reads the loaded picture and feeds `Size preset = Custom` plus that exact width and height, so the edit keeps the source resolution and aspect ratio.\n\nTo use a recommended size instead:\n1. Delete the `Get Image Size` node (its width/height links disappear with it).\n2. Pick a preset in `Size preset`, e.g. `(1K) 1024x1024 (1:1)`, `(1.5K) 1328x1776 (3:4)`, `(2K) 2048x2048 (1:1)`.\n3. Or keep `Custom` and type `width` / `height` yourself.\n\nA preset re-composes the canvas at that size; the edit still follows the source picture.\n"
301 },
302 "color": "#222",
303 "bgcolor": "#000"
304 }
305 ],
306 "links": [
307 [
308 18,
309 11,
310 0,
311 38,
312 0,
313 "IMAGE"
314 ],
315 [
316 19,
317 32,
318 0,
319 38,
320 1,
321 "IMAGE"
322 ],
323 [
324 21,
325 40,
326 0,
327 32,
328 0,
329 "IMAGE"
330 ],
331 [
332 22,
333 11,
334 0,
335 41,
336 0,
337 "IMAGE"
338 ],
339 [
340 23,
341 41,
342 0,
343 40,
344 0,
345 "INT"
346 ],
347 [
348 24,
349 41,
350 1,
351 40,
352 1,
353 "INT"
354 ],
355 [
356 20,
357 11,
358 0,
359 40,
360 2,
361 "IMAGE"
362 ]
363 ],
364 "groups": [],
365 "config": [],
366 "extra": {
367 "ds": {
368 "scale": 0.5730855330116845,
369 "offset": [
370 710.770457933314,
371 504.37190843288073
372 ]
373 },
374 "frontendVersion": "1.58.5"
375 },
376 "version": 0.4
377}