ComfyPilotComfyPilot
工作流 / 视频 / Wan3.0：参考生视频
Wan3.0：参考生视频

Wan3.0：参考生视频

Wan3.0: Reference to Video

使用Wan3.0的参考生视频功能，从两张参考图像生成视频，可生成长达30秒的片段，保持角色一致性并呈现逼真视觉效果。非常适合创建一致的角色动画、从视觉参考中复制场景，以及从静态图像制作视听内容。

视频 5 个节点 2 个输入 666 次使用 ComfyUI ≥ 0.33.4 ComfyUI 13 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

LoadImage × 2
MarkdownNote × 1
Wan3ReferenceToVideoApi × 1
SaveVideo × 1

官方标注的模型

Wan3.0

工作流 JSON

和上面「下载工作流 JSON」是同一份文件。

1{
2 "id": "9a8da4c0-0c4f-4a2f-a741-31e3c442f533",
3 "revision": 0,
4 "last_node_id": 18,
5 "last_link_id": 13,
6 "nodes": [
7 {
8 "id": 6,
9 "type": "LoadImage",
10 "pos": [
11 -3030,
12 5410
13 ],
14 "size": [
15 290,
16 350
17 ],
18 "flags": [],
19 "order": 0,
20 "mode": 0,
21 "inputs": [],
22 "outputs": [
23 {
24 "name": "IMAGE",
25 "type": "IMAGE",
26 "links": [
27 12
28 ]
29 },
30 {
31 "name": "MASK",
32 "type": "MASK",
33 "links": null
34 }
35 ],
36 "properties": {
37 "cnr_id": "comfy-core",
38 "ver": "0.33.0",
39 "Node name for S&R": "LoadImage"
40 },
41 "widgets_values": [
42 "vortex_shoe.png",
43 "image"
44 ],
45 "widgets_values_named": {
46 "image": "vortex_shoe.png",
47 "upload": "image"
48 }
49 },
50 {
51 "id": 14,
52 "type": "MarkdownNote",
53 "pos": [
54 -3460,
55 5410
56 ],
57 "size": [
58 380,
59 400
60 ],
61 "flags": [],
62 "order": 1,
63 "mode": 0,
64 "inputs": [],
65 "outputs": [],
66 "title": "Note: Input Asset",
67 "properties": [],
68 "widgets_values": [
69 "**Input Asset** \n\n\n- [vortex_shoe.png](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/input/vortex_shoe.png) — required workflow input (`input/vortex_shoe.png`)\n- [psss_character.png](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/input/psss_character.png) — required workflow input (`input/psss_character.png`)\n\n\n## Report Issue\n\nNote: Please update ComfyUI first ([guide](https://docs.comfy.org/installation/update_comfyui)) and prepare required models. Desktop/Cloud updates follow stable releases, so some nightly-supported models may not be available yet.\n\n- Cannot run / runtime errors: [ComfyUI/issues](https://github.com/comfyanonymous/ComfyUI/issues)\n- UI / frontend issues: [ComfyUI_frontend/issues](https://github.com/Comfy-Org/ComfyUI_frontend/issues)\n- Workflow issues: [workflow_templates/issues](https://github.com/Comfy-Org/workflow_templates/issues)\n"
70 ],
71 "widgets_values_named": {
72 "text": "**Input Asset** \n\n\n- [vortex_shoe.png](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/input/vortex_shoe.png) — required workflow input (`input/vortex_shoe.png`)\n- [psss_character.png](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/input/psss_character.png) — required workflow input (`input/psss_character.png`)\n\n\n## Report Issue\n\nNote: Please update ComfyUI first ([guide](https://docs.comfy.org/installation/update_comfyui)) and prepare required models. Desktop/Cloud updates follow stable releases, so some nightly-supported models may not be available yet.\n\n- Cannot run / runtime errors: [ComfyUI/issues](https://github.com/comfyanonymous/ComfyUI/issues)\n- UI / frontend issues: [ComfyUI_frontend/issues](https://github.com/Comfy-Org/ComfyUI_frontend/issues)\n- Workflow issues: [workflow_templates/issues](https://github.com/Comfy-Org/workflow_templates/issues)\n"
73 },
74 "color": "#222",
75 "bgcolor": "#000"
76 },
77 {
78 "id": 16,
79 "type": "Wan3ReferenceToVideoApi",
80 "pos": [
81 -2690,
82 5410
83 ],
84 "size": [
85 400,
86 390
87 ],
88 "flags": [],
89 "order": 3,
90 "mode": 0,
91 "inputs": [
92 {
93 "label": "image1",
94 "name": "model.reference_images.image1",
95 "shape": 7,
96 "type": "IMAGE",
97 "link": 12
98 },
99 {
100 "label": "image2",
101 "name": "model.reference_images.image2",
102 "shape": 7,
103 "type": "IMAGE",
104 "link": 13
105 },
106 {
107 "label": "image3",
108 "name": "model.reference_images.image3",
109 "shape": 7,
110 "type": "IMAGE",
111 "link": null
112 },
113 {
114 "label": "video1",
115 "name": "model.reference_videos.video1",
116 "shape": 7,
117 "type": "VIDEO",
118 "link": null
119 },
120 {
121 "label": "audio1",
122 "name": "model.reference_audios.audio1",
123 "shape": 7,
124 "type": "AUDIO",
125 "link": null
126 }
127 ],
128 "outputs": [
129 {
130 "name": "VIDEO",
131 "type": "VIDEO",
132 "links": [
133 11
134 ]
135 }
136 ],
137 "properties": {
138 "cnr_id": "comfy-core",
139 "ver": "0.33.0",
140 "Node name for S&R": "Wan3ReferenceToVideoApi"
141 },
142 "widgets_values": [
143 "wan3.0-video",
144 "Generate a 5-second, 16:9 street dance shoe showcase video: using @Image1 as the shoe reference and @Image2 as the character reference, a young man wearing the orange-and-white sneakers performs a street dance routine in front of a deep-blue backdrop, showcasing the shoes in motion throughout.\n\nMultimodal reference layer:\n- @Image1 (product spec): strictly keep the shoe design: the orange-and-white color blocking, streamlined silhouette, engineered mesh upper, breathable textile lining, dual-density midsole foam, Max-Air unit at the heel, and dynamic lacing system; ignore the callout text, arrows, numbering, and exploded view in the original\n- @Image2 (character portrait): strictly keep the character and the backdrop tone: a young man with short dark hair, a plain grey hoodie, futuristic white-framed sunglasses with blue iridescent lenses, on a deep-blue background; no text anywhere in the frame\n\nGlobal settings:\n- Scene and materials: seamless deep-blue studio backdrop matching the character reference, subtle floor reflections; streetwear commercial texture, high saturation and contrast\n- Visual style: hip-hop dance commercial, dynamic composition with a strong edit feel, the orange-and-white shoe as the visual anchor against the deep blue\n- Camera language: opening tilt-down to the feet, low-angle follow on the footwork, closing slide-and-freeze, the camera glued to the shoe's movement the whole time\n- Lighting: hard top light plus side-backlight with a cool rim glow, the floor reflecting the shoe's motion trail\n\nStoryboard:\n[0-1.5s] Slide-in entry: the character glides sideways into frame from off-screen, body grooving to the beat, the camera whipping down from a half-body shot to the feet, the orange-and-white sneaker drawing a smooth slide across the floor, sole squeaking softly\n\n[1.5-3.5s] Dance combo: the character chains a wave, kick, spin, and squat in a continuous street dance combo, the camera following the footwork low-angle as mesh texture, dynamic lacing, and the translucent air unit flash clearly in the fast motion, every landing hitting the floor with a crisp thud\n[3.5-5s] Hero freeze: the character slides on one knee to the camera, leaning back with one finger pointed at the lens; the camera finishes on a close-up of the shoe, the orange-and-white sneaker stopping dead center, freezing into a dynamic hero shot\n\nAudio design: layered ambience: heavy-bass hip-hop beat with DJ scratches; sole friction, shoe landings, and fabric whip sounds, every move landing on the beat; no dialogue, no voiceover.\n\nNegative constraints: hands never in pockets; no text, graffiti, callouts, or UI anywhere in the frame; no subtitles, no black frames, no jump cuts, no camera shake; do not alter the shoe's colorway or the character's outfit.",
145 "720P",
146 "adaptive",
147 "auto",
148 true,
149 true,
150 1341870941,
151 "randomize",
152 false
153 ],
154 "widgets_values_named": {
155 "model": "wan3.0-video",
156 "model.prompt": "Generate a 5-second, 16:9 street dance shoe showcase video: using @Image1 as the shoe reference and @Image2 as the character reference, a young man wearing the orange-and-white sneakers performs a street dance routine in front of a deep-blue backdrop, showcasing the shoes in motion throughout.\n\nMultimodal reference layer:\n- @Image1 (product spec): strictly keep the shoe design: the orange-and-white color blocking, streamlined silhouette, engineered mesh upper, breathable textile lining, dual-density midsole foam, Max-Air unit at the heel, and dynamic lacing system; ignore the callout text, arrows, numbering, and exploded view in the original\n- @Image2 (character portrait): strictly keep the character and the backdrop tone: a young man with short dark hair, a plain grey hoodie, futuristic white-framed sunglasses with blue iridescent lenses, on a deep-blue background; no text anywhere in the frame\n\nGlobal settings:\n- Scene and materials: seamless deep-blue studio backdrop matching the character reference, subtle floor reflections; streetwear commercial texture, high saturation and contrast\n- Visual style: hip-hop dance commercial, dynamic composition with a strong edit feel, the orange-and-white shoe as the visual anchor against the deep blue\n- Camera language: opening tilt-down to the feet, low-angle follow on the footwork, closing slide-and-freeze, the camera glued to the shoe's movement the whole time\n- Lighting: hard top light plus side-backlight with a cool rim glow, the floor reflecting the shoe's motion trail\n\nStoryboard:\n[0-1.5s] Slide-in entry: the character glides sideways into frame from off-screen, body grooving to the beat, the camera whipping down from a half-body shot to the feet, the orange-and-white sneaker drawing a smooth slide across the floor, sole squeaking softly\n\n[1.5-3.5s] Dance combo: the character chains a wave, kick, spin, and squat in a continuous street dance combo, the camera following the footwork low-angle as mesh texture, dynamic lacing, and the translucent air unit flash clearly in the fast motion, every landing hitting the floor with a crisp thud\n[3.5-5s] Hero freeze: the character slides on one knee to the camera, leaning back with one finger pointed at the lens; the camera finishes on a close-up of the shoe, the orange-and-white sneaker stopping dead center, freezing into a dynamic hero shot\n\nAudio design: layered ambience: heavy-bass hip-hop beat with DJ scratches; sole friction, shoe landings, and fabric whip sounds, every move landing on the beat; no dialogue, no voiceover.\n\nNegative constraints: hands never in pockets; no text, graffiti, callouts, or UI anywhere in the frame; no subtitles, no black frames, no jump cuts, no camera shake; do not alter the shoe's colorway or the character's outfit.",
157 "model.resolution": "720P",
158 "model.ratio": "adaptive",
159 "model.duration": "auto",
160 "model.audio": true,
161 "model.prompt_extend": true,
162 "seed": 1341870941,
163 "control_after_generate": "randomize",
164 "watermark": false
165 },
166 "color": "#432",
167 "bgcolor": "#653"
168 },
169 {
170 "id": 17,
171 "type": "SaveVideo",
172 "pos": [
173 -2240,
174 5410
175 ],
176 "size": [
177 620,
178 125.90625
179 ],
180 "flags": [],
181 "order": 4,
182 "mode": 0,
183 "inputs": [
184 {
185 "name": "video",
186 "type": "VIDEO",
187 "link": 11
188 }
189 ],
190 "outputs": [
191 {
192 "name": "video",
193 "type": "VIDEO",
194 "links": null
195 }
196 ],
197 "properties": {
198 "cnr_id": "comfy-core",
199 "ver": "0.33.0"
200 },
201 "widgets_values": [
202 "video/Wan3.0_r2v",
203 "auto",
204 "auto",
205 "auto"
206 ],
207 "widgets_values_named": {
208 "filename_prefix": "video/Wan3.0_r2v",
209 "format": "auto",
210 "format.codec": "auto",
211 "codec": "auto"
212 }
213 },
214 {
215 "id": 18,
216 "type": "LoadImage",
217 "pos": [
218 -3030,
219 5840
220 ],
221 "size": [
222 290,
223 350
224 ],
225 "flags": [],
226 "order": 2,
227 "mode": 0,
228 "inputs": [],
229 "outputs": [
230 {
231 "name": "IMAGE",
232 "type": "IMAGE",
233 "links": [
234 13
235 ]
236 },
237 {
238 "name": "MASK",
239 "type": "MASK",
240 "links": null
241 }
242 ],
243 "properties": {
244 "cnr_id": "comfy-core",
245 "ver": "0.33.0",
246 "Node name for S&R": "LoadImage"
247 },
248 "widgets_values": [
249 "psss_character.png",
250 "image"
251 ],
252 "widgets_values_named": {
253 "image": "psss_character.png",
254 "upload": "image"
255 }
256 }
257 ],
258 "links": [
259 [
260 11,
261 16,
262 0,
263 17,
264 0,
265 "VIDEO"
266 ],
267 [
268 12,
269 6,
270 0,
271 16,
272 0,
273 "IMAGE"
274 ],
275 [
276 13,
277 18,
278 0,
279 16,
280 1,
281 "IMAGE"
282 ]
283 ],
284 "groups": [],
285 "config": [],
286 "extra": {
287 "ds": {
288 "scale": 0.1,
289 "offset": [
290 10066.884762777934,
291 -1592.7250014845395
292 ]
293 },
294 "frontendVersion": "1.49.6",
295 "VHS_latentpreview": false,
296 "VHS_latentpreviewrate": 0,
297 "VHS_MetadataImage": true,
298 "VHS_KeepIntermediate": true
299 },
300 "version": 0.4
301}