工作流 /
视频 /
HeyGen Avatar 6:照片生视频
HeyGen Avatar 6:照片生视频
HeyGen Avatar 6: Photo to Video
一张照片加一段配音,生成会说话的视频。HeyGen Avatar 6 保留照片中的人物与场景,让口型跟随你的音频,并按动作提示词控制手势、表情、姿态与镜头。视频长度等于音频长度,可选 1080p 或 2K,引擎可选 Avatar 6 或更便宜的 Avatar 6 Turbo。适合让人像开口说话、无需拍摄制作出镜口播,或让角色插画活起来。
视频
7 个节点 2 个输入 ComfyUI ≥ 0.46.0 ComfyUI 10 KB
节点构成
MarkdownNote
× 2
LoadImage
× 1
LoadAudio
× 1
HeyGenAvatar6ReferenceToVideoNode
× 1
SaveVideo
× 1
HeyGenTextToSpeechNode
× 1
官方标注的模型
HeyGen Avatar 6
工作流 JSON
和上面「下载工作流 JSON」是同一份文件。
1{
2 "id": "a27869ac-3e26-4094-a1af-a1b45c55f0d6",
3 "revision": 0,
4 "last_node_id": 8,
5 "last_link_id": 3,
6 "nodes": [
7 {
8 "id": 1,
9 "type": "LoadImage",
10 "pos": [
11 -386.11122530652904,
12 514.765322411633
13 ],
14 "size": [
15 456.73468094482433,
16 485.5845932952882
17 ],
18 "flags": [],
19 "order": 0,
20 "mode": 0,
21 "inputs": [],
22 "outputs": [
23 {
24 "name": "IMAGE",
25 "type": "IMAGE",
26 "links": [
27 2
28 ]
29 },
30 {
31 "name": "MASK",
32 "type": "MASK",
33 "links": null
34 }
35 ],
36 "properties": {
37 "Node name for S&R": "LoadImage"
38 },
39 "widgets_values": [
40 "pink_hair_portrait.png",
41 "image"
42 ],
43 "widgets_values_named": {
44 "image": "pink_hair_portrait.png",
45 "upload": "image"
46 }
47 },
48 {
49 "id": 2,
50 "type": "LoadAudio",
51 "pos": [
52 -386.11122530652904,
53 310.9859171595023
54 ],
55 "size": [
56 459.2990389737215,
57 138.00004438920453
58 ],
59 "flags": [],
60 "order": 1,
61 "mode": 0,
62 "inputs": [],
63 "outputs": [
64 {
65 "name": "AUDIO",
66 "type": "AUDIO",
67 "links": [
68 1
69 ]
70 }
71 ],
72 "properties": {
73 "Node name for S&R": "LoadAudio"
74 },
75 "widgets_values": [
76 "avatar6_narration.mp3",
77 ""
78 ],
79 "widgets_values_named": {
80 "audio": "avatar6_narration.mp3",
81 "upload": ""
82 }
83 },
84 {
85 "id": 3,
86 "type": "HeyGenAvatar6ReferenceToVideoNode",
87 "pos": [
88 121.18781366719247,
89 310.9859171595023
90 ],
91 "size": [
92 400,
93 296
94 ],
95 "flags": [],
96 "order": 5,
97 "mode": 0,
98 "inputs": [
99 {
100 "name": "model.audio",
101 "type": "AUDIO",
102 "link": 1
103 },
104 {
105 "name": "model.image",
106 "shape": 7,
107 "type": "IMAGE",
108 "link": 2
109 },
110 {
111 "name": "model.video",
112 "shape": 7,
113 "type": "VIDEO",
114 "link": null
115 }
116 ],
117 "outputs": [
118 {
119 "name": "VIDEO",
120 "type": "VIDEO",
121 "links": [
122 3
123 ]
124 }
125 ],
126 "properties": {
127 "Node name for S&R": "HeyGenAvatar6ReferenceToVideoNode"
128 },
129 "widgets_values": [
130 "heygen-avatar-6",
131 "1080p",
132 "auto",
133 "She speaks straight to camera with calm confidence, small open-handed gestures on the key words, relaxed shoulders and a slight head tilt; the camera eases into a slow, subtle push-in.",
134 "default",
135 1558795324,
136 "randomize",
137 "randomize"
138 ],
139 "widgets_values_named": {
140 "model": "heygen-avatar-6",
141 "model.resolution": "1080p",
142 "model.aspect_ratio": "auto",
143 "model.motion_prompt": "She speaks straight to camera with calm confidence, small open-handed gestures on the key words, relaxed shoulders and a slight head tilt; the camera eases into a slow, subtle push-in.",
144 "model.motion_expressiveness": "default",
145 "model.seed": 1558795324,
146 "control_after_generate#1": "randomize",
147 "control_after_generate": "randomize"
148 },
149 "color": "#432",
150 "bgcolor": "#653"
151 },
152 {
153 "id": 4,
154 "type": "SaveVideo",
155 "pos": [
156 569.1878136671925,
157 310.9859171595023
158 ],
159 "size": [
160 565.5803210449219,
161 500.45856445312506
162 ],
163 "flags": [],
164 "order": 6,
165 "mode": 0,
166 "inputs": [
167 {
168 "name": "video",
169 "type": "VIDEO",
170 "link": 3
171 }
172 ],
173 "outputs": [
174 {
175 "name": "video",
176 "type": "VIDEO",
177 "links": null
178 }
179 ],
180 "properties": [],
181 "widgets_values": [
182 "video/HeyGen_avatar6_ia2v",
183 "auto",
184 "auto",
185 "auto"
186 ],
187 "widgets_values_named": {
188 "filename_prefix": "video/HeyGen_avatar6_ia2v",
189 "format": "auto",
190 "format.codec": "auto",
191 "codec": "auto"
192 }
193 },
194 {
195 "id": 5,
196 "type": "MarkdownNote",
197 "pos": [
198 -967.7464264428925,
199 316.17718239387716
200 ],
201 "size": [
202 509.6815912353517,
203 627.620031567383
204 ],
205 "flags": [],
206 "order": 2,
207 "mode": 0,
208 "inputs": [],
209 "outputs": [],
210 "properties": [],
211 "widgets_values": [
212 "## HeyGen Avatar 6: Photo to Video\n\nOne photo plus one voice track becomes a talking clip. The photo carries the identity and the\nscene, the AUDIO clip sets the length, and a motion prompt directs the performance.\n\n## Input Assets\n\n- [pink_hair_portrait.png](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/input/pink_hair_portrait.png)\n- [avatar6_narration.mp3](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/input/avatar6_narration.mp3)\n\n### How to use\n\n1. Sign in with your ComfyUI account. This is an API node, it bills per render.\n2. Load a head-and-shoulders or half-body photo with the hands free — the model re-stages the\n pose, so a photo with the arms crossed or the hands on the face comes back re-posed.\n3. Connect an AUDIO clip (1.5 s to 10 minutes). The clip length is the render length.\n4. `model` picks the engine (`heygen-avatar-6` or the cheaper `heygen-avatar-6-turbo`);\n `model.resolution` picks 1080p or 2k and `model.aspect_ratio` = `auto` keeps the photo's shape.\n5. `model.motion_prompt` directs gestures, expression, posture and camera movement (up to 1200\n characters). `model.motion_expressiveness` = `less` damps the gestures.\n6. The photo is not used as the first frame: frame one is re-staged from the photo.\n\n### What to expect\n\n- Video of **realistic people in a reference image** is fine; only reference *videos* of realistic\n people are refused (that is the sibling reference-to-video template).\n- Output size follows the photo aspect: this 1024x1024 photo comes back 1088x1088.\n- `seed` only decides whether the node re-runs; renders are non-deterministic anyway.\n"
213 ],
214 "widgets_values_named": {
215 "text": "## HeyGen Avatar 6: Photo to Video\n\nOne photo plus one voice track becomes a talking clip. The photo carries the identity and the\nscene, the AUDIO clip sets the length, and a motion prompt directs the performance.\n\n## Input Assets\n\n- [pink_hair_portrait.png](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/input/pink_hair_portrait.png)\n- [avatar6_narration.mp3](https://raw.githubusercontent.com/Comfy-Org/workflow_templates/refs/heads/main/input/avatar6_narration.mp3)\n\n### How to use\n\n1. Sign in with your ComfyUI account. This is an API node, it bills per render.\n2. Load a head-and-shoulders or half-body photo with the hands free — the model re-stages the\n pose, so a photo with the arms crossed or the hands on the face comes back re-posed.\n3. Connect an AUDIO clip (1.5 s to 10 minutes). The clip length is the render length.\n4. `model` picks the engine (`heygen-avatar-6` or the cheaper `heygen-avatar-6-turbo`);\n `model.resolution` picks 1080p or 2k and `model.aspect_ratio` = `auto` keeps the photo's shape.\n5. `model.motion_prompt` directs gestures, expression, posture and camera movement (up to 1200\n characters). `model.motion_expressiveness` = `less` damps the gestures.\n6. The photo is not used as the first frame: frame one is re-staged from the photo.\n\n### What to expect\n\n- Video of **realistic people in a reference image** is fine; only reference *videos* of realistic\n people are refused (that is the sibling reference-to-video template).\n- Output size follows the photo aspect: this 1024x1024 photo comes back 1088x1088.\n- `seed` only decides whether the node re-runs; renders are non-deterministic anyway.\n"
216 },
217 "color": "#222",
218 "bgcolor": "#000"
219 },
220 {
221 "id": 7,
222 "type": "HeyGenTextToSpeechNode",
223 "pos": [
224 -337.10968071465453,
225 1158.093911413927
226 ],
227 "size": [
228 400,
229 232
230 ],
231 "flags": [],
232 "order": 3,
233 "mode": 4,
234 "inputs": [],
235 "outputs": [
236 {
237 "name": "AUDIO",
238 "type": "AUDIO",
239 "links": null
240 }
241 ],
242 "properties": {
243 "Node name for S&R": "HeyGenTextToSpeechNode"
244 },
245 "widgets_values": [
246 "",
247 "Chill Brian (English, male)",
248 "",
249 1,
250 false,
251 42,
252 "randomize"
253 ],
254 "widgets_values_named": {
255 "text": "",
256 "voice": "Chill Brian (English, male)",
257 "custom_voice_id": "",
258 "speed": 1,
259 "ssml": false,
260 "seed": 42,
261 "control_after_generate": "randomize"
262 },
263 "color": "#432",
264 "bgcolor": "#653"
265 },
266 {
267 "id": 8,
268 "type": "MarkdownNote",
269 "pos": [
270 -332.74862602715416,
271 1458.015083288925
272 ],
273 "size": [
274 394.87890625,
275 78
276 ],
277 "flags": [],
278 "order": 4,
279 "mode": 0,
280 "inputs": [],
281 "outputs": [],
282 "properties": [],
283 "widgets_values": [
284 "Select and Ctrl-B to enable, then plug the `AUDIO` output to the audio input of the `HeyGen Avatar 6 Reference to Video` node"
285 ],
286 "widgets_values_named": {
287 "text": "Select and Ctrl-B to enable, then plug the `AUDIO` output to the audio input of the `HeyGen Avatar 6 Reference to Video` node"
288 },
289 "color": "#222",
290 "bgcolor": "#000"
291 }
292 ],
293 "links": [
294 [
295 1,
296 2,
297 0,
298 3,
299 0,
300 "AUDIO"
301 ],
302 [
303 2,
304 1,
305 0,
306 3,
307 1,
308 "IMAGE"
309 ],
310 [
311 3,
312 3,
313 0,
314 4,
315 0,
316 "VIDEO"
317 ]
318 ],
319 "groups": [
320 {
321 "id": 1,
322 "title": "Optional (Create audio for audio input)",
323 "bounding": [
324 -392.8842900896545,
325 1060.8259426639268,
326 509.30976562499995,
327 485.3089843749999
328 ],
329 "color": "#3f789e",
330 "flags": []
331 }
332 ],
333 "config": [],
334 "extra": {
335 "ds": {
336 "scale": 0.35049389948139237,
337 "offset": [
338 2986.317520639524,
339 555.7913379324497
340 ]
341 },
342 "frontendVersion": "1.59.1"
343 },
344 "version": 0.4
345}