工作流 /
音频 /
种子音频 1.0：文生音频
种子音频 1.0：文生音频
Seed Audio 1.0: Text to Audio
输入文本提示词以生成语音、对话、背景音乐和声音效果，并合并到一个音频文件中。描述声音、情感和场景详情，以创建最多2分钟的多说话人音频。
音频
3 个节点 67 次使用 ComfyUI ≥ 0.27.1 ComfyUI 6 KB
节点构成
ByteDanceSeedAudio
× 1
SaveAudioAdvanced
× 1
MarkdownNote
× 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": 5,
5 "last_link_id": 2,
6 "nodes": [
7 {
8 "id": 3,
9 "type": "ByteDanceSeedAudio",
10 "pos": [
11 -261.01158597793875,
12 -384.6790567581335
13 ],
14 "size": [
15 495.390625,
16 710.609375
17 ],
18 "flags": [],
19 "order": 0,
20 "mode": 0,
21 "inputs": [],
22 "outputs": [
23 {
24 "name": "AUDIO",
25 "type": "AUDIO",
26 "links": [
27 2
28 ]
29 }
30 ],
31 "properties": {
32 "Node name for S&R": "ByteDanceSeedAudio"
33 },
34 "widgets_values": [
35 "A quiet coffee shop on a rainy afternoon, soft jazz playing from an old speaker, rain tapping against the window, steam rising from a ceramic cup.\n\nLeo (young man, tired but warm voice, slightly hoarse, American accent) says slowly: \"I didn't think you'd actually show up.\"\n\nNina (young woman, soft and careful voice, a faint smile in her tone) replies gently: \"Neither did I. But here we are.\"\n\n[Long pause. The sound of stirring a spoon against a cup.]\n\nLeo, quieter now, almost hesitant: \"Do you ever wonder what would've happened if we'd met five years earlier?\"\n\nNina lets out a small, sad laugh: \"I used to. Now I just wonder what happens tomorrow.\"\n\n[Brief silence. Rain grows slightly louder. Jazz piano swells for a moment.]\n\nLeo, voice cracking slightly: \"I wish I had a better answer for you.\"\n\nNina, soft and warm: \"You being here is answer enough.\"\n\n[Music fades slowly. Rain continues. Their breathing, then silence.]",
36 "text only",
37 "24000",
38 0,
39 0,
40 0,
41 367178397,
42 "randomize"
43 ],
44 "color": "#432",
45 "bgcolor": "#653"
46 },
47 {
48 "id": 4,
49 "type": "SaveAudioAdvanced",
50 "pos": [
51 275.74282959116664,
52 -384.8495210264142
53 ],
54 "size": [
55 621.46875,
56 176
57 ],
58 "flags": [],
59 "order": 2,
60 "mode": 0,
61 "inputs": [
62 {
63 "name": "audio",
64 "type": "AUDIO",
65 "link": 2
66 }
67 ],
68 "outputs": [
69 {
70 "name": "audio",
71 "type": "AUDIO",
72 "links": null
73 }
74 ],
75 "properties": [],
76 "widgets_values": [
77 "audio/Seed_Audio_1.0_t2a",
78 "mp3",
79 "V0"
80 ]
81 },
82 {
83 "id": 5,
84 "type": "MarkdownNote",
85 "pos": [
86 -776.6736744605338,
87 -388.55370420891586
88 ],
89 "size": [
90 470.21875,
91 1704.109375
92 ],
93 "flags": [],
94 "order": 1,
95 "mode": 0,
96 "inputs": [],
97 "outputs": [],
98 "title": "Note: T2A Prompt Guide",
99 "properties": [],
100 "widgets_values": [
101 "Seed Audio 1.0: T2A (Text to Audio) Workflow Guide\n\n## What This Workflow Does\nByteDance Seed Audio 1.0 (text only) → Save Audio.\n\nT2A = text prompt only. No reference image or audio. Describe the full scene in one prompt: voices, dialogue, ambience, background music, and sound effects. The model generates them together.\n\nBest for: quick scenes, multi-character dialogue, audiobook clips, dubbing when you can describe each voice in text. Also use this workflow to create a clean reference clip for the TA2A workflow.\n\n## How to Use\n1. Open ByteDance Seed Audio 1.0 and keep reference_mode on \"text only\".\n2. Write your scene in text_prompt (see format below).\n3. Run the workflow. Output saves via Save Audio.\n\nNo extra inputs required. Do 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 at the top:\n- [Language: English only.]\n- [Language: Chinese only.]\n\nAlso state the language in each character's voice traits, for example: speaks English only, or speaks Chinese only.\n\n## Avoid Scene Text Being Spoken\nPlain prose without brackets 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- [SFX: ...] or [Long pause. ...] for effects and stage directions\n- [Outro: ...] for ending\n\nOnly text in quotes after says / replies / whispers should be spoken.\n\n## Prompt Guide\n1. [Language: ...] optional but recommended\n2. [Environment: ...]\n3. [Background music / SFX: ...]\n4. Character voice (gender, age, accent, emotion, tone, pace, target language)\n5. Exact dialogue in quotes\n6. [Beats / outro]\n\nLimits: prompt ≤3000 chars · output ≤2 min · English & Chinese\n\n## Example\n[Language: English only.]\n[Environment: quiet coffee shop on a rainy afternoon, rain on window, soft room tone.]\n[Background music: soft jazz from an old speaker.]\n\nLeo (young man, tired but warm voice, American accent, speaks English only) says slowly:\n\"I didn't think you'd actually show up.\"\n\nNina (young woman, soft careful voice, speaks English only) replies gently:\n\"Neither did I. But here we are.\"\n\n[Long pause. Spoon stirring a cup. Rain grows louder. Jazz piano swells.]\n[Music fades. Rain continues. Silence.]"
102 ],
103 "color": "#222",
104 "bgcolor": "#000"
105 }
106 ],
107 "links": [
108 [
109 2,
110 3,
111 0,
112 4,
113 0,
114 "AUDIO"
115 ]
116 ],
117 "groups": [],
118 "config": [],
119 "extra": {
120 "ds": {
121 "scale": 0.7226557052385247,
122 "offset": [
123 1005.0798545529957,
124 579.8527643179303
125 ]
126 },
127 "frontendVersion": "1.45.20"
128 },
129 "version": 0.4
130}