工作流 /
音频 /
种子音频 1.0：文本 + 图像转音频
种子音频 1.0：文本 + 图像转音频
Seed Audio 1.0: Text + Image to Audio
上传一张角色图像并编写文本提示词以生成音频。模型从您的图像中提取声音，然后生成语音、环境音、背景音乐和音效。
音频
4 个节点 1 个输入 13 次使用 ComfyUI ≥ 0.27.1 ComfyUI 6 KB
节点构成
ByteDanceSeedAudio
× 1
SaveAudioAdvanced
× 1
MarkdownNote
× 1
LoadImage
× 1
官方标注的模型
Seed Audio 1.0
工作流 JSON
和上面「下载工作流 JSON」是同一份文件。
1{
2 "id": "6d029989-7e4b-4077-acb4-e5c45eec2b1d",
3 "revision": 0,
4 "last_node_id": 6,
5 "last_link_id": 4,
6 "nodes": [
7 {
8 "id": 3,
9 "type": "ByteDanceSeedAudio",
10 "pos": [
11 -243.23859960711826,
12 -386.98716660616896
13 ],
14 "size": [
15 519.984375,
16 709
17 ],
18 "flags": [],
19 "order": 2,
20 "mode": 0,
21 "inputs": [
22 {
23 "label": "reference_image",
24 "name": "reference_mode.reference_image",
25 "shape": 7,
26 "type": "IMAGE",
27 "link": 4
28 }
29 ],
30 "outputs": [
31 {
32 "name": "AUDIO",
33 "type": "AUDIO",
34 "links": [
35 2
36 ]
37 }
38 ],
39 "properties": {
40 "Node name for S&R": "ByteDanceSeedAudio"
41 },
42 "widgets_values": [
43 "[Environment: quiet dreamlike garden at dusk, soft blue light, petals drifting, faint water ripples, goldfish swimming overhead.]\n\n[Background music: gentle piano and airy strings, slow, melancholic, ethereal. Ambience: soft wind, distant chimes.]\n\n[Language: English only. All dialogue must be spoken in English.]\n\nThe girl (young woman, soft breathy voice, quiet and fragile, dreamy, slightly sad) whispers:\n\"I thought the flowers would stay with me forever.\"\n\n[Long pause. Petals fall. Water shimmers.]\n\nShe (same voice, a little warmer) says softly:\n\"Maybe that is why this place feels beautiful.\"\n\n[Background music fades to a single piano note, then silence.]",
44 "image reference",
45 "24000",
46 0,
47 0,
48 0,
49 886100183,
50 "randomize"
51 ],
52 "color": "#432",
53 "bgcolor": "#653"
54 },
55 {
56 "id": 4,
57 "type": "SaveAudioAdvanced",
58 "pos": [
59 345.59583883312115,
60 -387.3237445260305
61 ],
62 "size": [
63 644.453125,
64 216
65 ],
66 "flags": [],
67 "order": 3,
68 "mode": 0,
69 "inputs": [
70 {
71 "name": "audio",
72 "type": "AUDIO",
73 "link": 2
74 }
75 ],
76 "outputs": [
77 {
78 "name": "audio",
79 "type": "AUDIO",
80 "links": null
81 }
82 ],
83 "properties": [],
84 "widgets_values": [
85 "audio/Seed_Audio_1.0",
86 "mp3",
87 "V0"
88 ]
89 },
90 {
91 "id": 5,
92 "type": "MarkdownNote",
93 "pos": [
94 -1192.5262732565643,
95 -389.08180319438
96 ],
97 "size": [
98 470.21875,
99 1521.71875
100 ],
101 "flags": [],
102 "order": 0,
103 "mode": 0,
104 "inputs": [],
105 "outputs": [],
106 "title": "Note: TI2A Prompt Guide",
107 "properties": [],
108 "widgets_values": [
109 "Seed Audio 1.0: TI2A (Text + Image to Audio) Workflow Guide\n\n## What This Workflow Does\nLoad Image → ByteDance Seed Audio 1.0 (image reference) → Save Audio.\n\nTI2A = text prompt + reference image. The model derives the character voice from your image, then generates speech, ambience, music, and SFX from your prompt.\n\nBest for: character art, portraits, or scene images when you have no reference audio. One image = one derived voice.\n\n## How to Use\n1. Load Image: upload a clear character image (portrait or upper body works best).\n2. Connect IMAGE → Seed Audio reference_image input.\n3. Keep reference_mode on \"image reference\".\n4. Write text_prompt matching your scene (see format below).\n5. Run the workflow.\n\nDo NOT use @Audio1 / @Audio2 / @Audio3 in this mode.\n\n## Target Language\nThe model supports English and Chinese. It may mix languages if you do not specify.\n\nTo lock the output language, add a bracket tag near the top of your prompt:\n- [Language: English only.]\n- [Language: Chinese only.]\n\nAlso state the language in the character's voice traits, for example: speaks English only, or speaks Chinese only. Write all dialogue in your target language.\n\nImage reference controls voice tone and character feel. Language is controlled by your prompt.\n\n## Avoid Scene Text Being Spoken\nPlain prose at the start may be read aloud as narration, making the audio too long.\n\nWrap anything that is NOT dialogue in [square brackets]:\n- [Environment: ...] for location, weather, ambience\n- [Background music: ...] for BGM style and mood\n- [Language: English only.] or [Language: Chinese only.] for output language\n- [Long pause. ...] for effects and stage directions\n- [Outro: ...] for ending\n\nOnly text in quotes after says / whispers should be spoken. Keep dialogue short.\n\n## Prompt Guide\n1. [Language: ...] recommended\n2. [Environment: ...]\n3. [Background music / SFX: ...]\n4. Character (delivery traits, target language) says / whispers: \"dialogue only\"\n5. [Beat / pause / SFX]\n6. [Outro]\n\nLimits: prompt ≤3000 chars · output ≤2 min · English & Chinese\n\n## Example\n[Language: English only.]\n[Environment: quiet dreamlike garden at dusk, soft blue light, petals drifting, faint water ripples.]\n[Background music: gentle piano and airy strings, melancholic and ethereal.]\n\nThe girl (young woman, soft breathy voice, quiet and fragile, dreamy, slightly sad, speaks English only) whispers:\n\"I thought the flowers would stay with me forever.\"\n\n[Long pause. Petals fall. Water shimmers.]\n\nShe (same voice, a little warmer) says softly:\n\"Maybe that is why this place feels beautiful.\"\n\n[Background music fades to silence.]"
110 ],
111 "color": "#222",
112 "bgcolor": "#000"
113 },
114 {
115 "id": 6,
116 "type": "LoadImage",
117 "pos": [
118 -670,
119 -390
120 ],
121 "size": [
122 386.625,
123 535.40625
124 ],
125 "flags": [],
126 "order": 1,
127 "mode": 0,
128 "inputs": [],
129 "outputs": [
130 {
131 "name": "IMAGE",
132 "type": "IMAGE",
133 "links": [
134 4
135 ]
136 },
137 {
138 "name": "MASK",
139 "type": "MASK",
140 "links": null
141 }
142 ],
143 "properties": {
144 "Node name for S&R": "LoadImage"
145 },
146 "widgets_values": [
147 "girl_and_fish.png",
148 "image"
149 ]
150 }
151 ],
152 "links": [
153 [
154 2,
155 3,
156 0,
157 4,
158 0,
159 "AUDIO"
160 ],
161 [
162 4,
163 6,
164 0,
165 3,
166 0,
167 "IMAGE"
168 ]
169 ],
170 "groups": [],
171 "config": [],
172 "extra": {
173 "ds": {
174 "scale": 0.7651805044950032,
175 "offset": [
176 1163.5846393445972,
177 881.0401016124736
178 ]
179 },
180 "frontendVersion": "1.45.20"
181 },
182 "version": 0.4
183}