Seed Audio 1.0

Seed Audio 1.0 is a ByteDance speech generation model built for high-quality, highly controllable audio output. Public technical materials describe the underlying Seed-TTS family as supporting zero-shot text-to-speech, cross-lingual speech generation, emotion control, speaker adaptation, voice conversion, speech editing, and speed editing, making it well suited to narration, dubbing, character voice work, and other expressive speech workflows.

API Reference
INTEGRATE
Complete technical specification for integration
RequestResponse
Examples4
CODE
Ready-to-use code snippets for common workflows
Subsea Cable Podcast Cold Open
0:00
import { createClient } from '@runware/sdk'
const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()
const [result] = await client.run({
model: 'bytedance:seed-audio@1.0',
positivePrompt: 'Create a finished 30-second documentary podcast cold open. Begin with four seconds of dark, spacious deep-ocean ambience: low hydrophone rumble, distant metallic hull creaks, and two restrained sonar-like pings. Introduce a precise, quietly urgent narrator with a mature, gender-neutral voice and natural broadcast delivery. Exact narration: "Two thousand meters below the weather, the internet is no longer invisible. It is glass, steel, current—and a repair crew racing the tide. This is Below the Backbone, a field podcast about the people who keep subsea cables alive. In this episode: the midnight splice that reconnected an island." Underscore the narration with a minimal sub-bass pulse and sparse glassy electronic tones, evoking fiber-optic signals without becoming futuristic or musical theater. Let the tension build subtly beneath the final sentence, then end with a clean three-note sonic logo and a short underwater decay. Premium investigative-journalism mix, intimate centered voice, wide environmental sound, crisp consonants, no additional speech, no singing, no advertising tone.',
speech: {
volume: 8,
speed: -4,
pitch: -1
}
})import asyncio
import os
from runware import Runware
async def main():
async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
results = await client.run({
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a finished 30-second documentary podcast cold open. Begin with four seconds of dark, spacious deep-ocean ambience: low hydrophone rumble, distant metallic hull creaks, and two restrained sonar-like pings. Introduce a precise, quietly urgent narrator with a mature, gender-neutral voice and natural broadcast delivery. Exact narration: \"Two thousand meters below the weather, the internet is no longer invisible. It is glass, steel, current—and a repair crew racing the tide. This is Below the Backbone, a field podcast about the people who keep subsea cables alive. In this episode: the midnight splice that reconnected an island.\" Underscore the narration with a minimal sub-bass pulse and sparse glassy electronic tones, evoking fiber-optic signals without becoming futuristic or musical theater. Let the tension build subtly beneath the final sentence, then end with a clean three-note sonic logo and a short underwater decay. Premium investigative-journalism mix, intimate centered voice, wide environmental sound, crisp consonants, no additional speech, no singing, no advertising tone.",
"speech": {
"volume": 8,
"speed": -4,
"pitch": -1
}
})
asyncio.run(main())curl https://api.runware.ai/v1 \
-H "Authorization: Bearer $RUNWARE_API_KEY" \
-H "Content-Type: application/json" \
-d '[
{
"taskType": "audioInference",
"taskUUID": "6c1a142d-1466-4c64-a1c5-169704a523da",
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a finished 30-second documentary podcast cold open. Begin with four seconds of dark, spacious deep-ocean ambience: low hydrophone rumble, distant metallic hull creaks, and two restrained sonar-like pings. Introduce a precise, quietly urgent narrator with a mature, gender-neutral voice and natural broadcast delivery. Exact narration: \"Two thousand meters below the weather, the internet is no longer invisible. It is glass, steel, current—and a repair crew racing the tide. This is Below the Backbone, a field podcast about the people who keep subsea cables alive. In this episode: the midnight splice that reconnected an island.\" Underscore the narration with a minimal sub-bass pulse and sparse glassy electronic tones, evoking fiber-optic signals without becoming futuristic or musical theater. Let the tension build subtly beneath the final sentence, then end with a clean three-note sonic logo and a short underwater decay. Premium investigative-journalism mix, intimate centered voice, wide environmental sound, crisp consonants, no additional speech, no singing, no advertising tone.",
"speech": {
"volume": 8,
"speed": -4,
"pitch": -1
}
}
]'runware run bytedance:seed-audio@1.0 \
positivePrompt="Create a finished 30-second documentary podcast cold open. Begin with four seconds of dark, spacious deep-ocean ambience: low hydrophone rumble, distant metallic hull creaks, and two restrained sonar-like pings. Introduce a precise, quietly urgent narrator with a mature, gender-neutral voice and natural broadcast delivery. Exact narration: \"Two thousand meters below the weather, the internet is no longer invisible. It is glass, steel, current—and a repair crew racing the tide. This is Below the Backbone, a field podcast about the people who keep subsea cables alive. In this episode: the midnight splice that reconnected an island.\" Underscore the narration with a minimal sub-bass pulse and sparse glassy electronic tones, evoking fiber-optic signals without becoming futuristic or musical theater. Let the tension build subtly beneath the final sentence, then end with a clean three-note sonic logo and a short underwater decay. Premium investigative-journalism mix, intimate centered voice, wide environmental sound, crisp consonants, no additional speech, no singing, no advertising tone." \
speech.volume=8 \
speech.speed=-4 \
speech.pitch=-1{
"taskType": "audioInference",
"taskUUID": "6c1a142d-1466-4c64-a1c5-169704a523da",
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a finished 30-second documentary podcast cold open. Begin with four seconds of dark, spacious deep-ocean ambience: low hydrophone rumble, distant metallic hull creaks, and two restrained sonar-like pings. Introduce a precise, quietly urgent narrator with a mature, gender-neutral voice and natural broadcast delivery. Exact narration: \"Two thousand meters below the weather, the internet is no longer invisible. It is glass, steel, current—and a repair crew racing the tide. This is Below the Backbone, a field podcast about the people who keep subsea cables alive. In this episode: the midnight splice that reconnected an island.\" Underscore the narration with a minimal sub-bass pulse and sparse glassy electronic tones, evoking fiber-optic signals without becoming futuristic or musical theater. Let the tension build subtly beneath the final sentence, then end with a clean three-note sonic logo and a short underwater decay. Premium investigative-journalism mix, intimate centered voice, wide environmental sound, crisp consonants, no additional speech, no singing, no advertising tone.",
"speech": {
"volume": 8,
"speed": -4,
"pitch": -1
}
}{
"taskType": "audioInference",
"taskUUID": "6c1a142d-1466-4c64-a1c5-169704a523da",
"audioUUID": "26e5e946-5565-41e0-9862-932daa461fd9",
"audioURL": "https://am.runware.ai/audio/os/a05d22/ws/5/ai/26e5e946-5565-41e0-9862-932daa461fd9.mp3",
"cost": 0.079
}Oyster Farm Forecast Radio Jingle
0:00
import { createClient } from '@runware/sdk'
const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()
const [result] = await client.run({
model: 'bytedance:seed-audio@1.0',
positivePrompt: 'Create a polished 10-second radio jingle for “The Oyster Line,” a daily marine forecast serving working oyster farms. Open with two soft buoy-bell notes and a subtle rolling-water sound, then introduce an upbeat coastal folk groove with muted mandolin, upright bass, brushed snare, and a clean sonar-like chime. A confident, friendly adult woman says rhythmically: “The Oyster Line — tides, wind, and water for the working coast.” End on a memorable three-note musical logo with a crisp, broadcast-safe finish. Warm, practical, contemporary, and trustworthy; no exaggerated pirate or nautical clichés. Clear speech centered above the music, professional radio mix, no long fade-out.',
speech: {
volume: 10,
speed: 8,
pitch: 0
}
})import asyncio
import os
from runware import Runware
async def main():
async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
results = await client.run({
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a polished 10-second radio jingle for “The Oyster Line,” a daily marine forecast serving working oyster farms. Open with two soft buoy-bell notes and a subtle rolling-water sound, then introduce an upbeat coastal folk groove with muted mandolin, upright bass, brushed snare, and a clean sonar-like chime. A confident, friendly adult woman says rhythmically: “The Oyster Line — tides, wind, and water for the working coast.” End on a memorable three-note musical logo with a crisp, broadcast-safe finish. Warm, practical, contemporary, and trustworthy; no exaggerated pirate or nautical clichés. Clear speech centered above the music, professional radio mix, no long fade-out.",
"speech": {
"volume": 10,
"speed": 8,
"pitch": 0
}
})
asyncio.run(main())curl https://api.runware.ai/v1 \
-H "Authorization: Bearer $RUNWARE_API_KEY" \
-H "Content-Type: application/json" \
-d '[
{
"taskType": "audioInference",
"taskUUID": "3c569a98-0158-4828-af18-43495bd51734",
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a polished 10-second radio jingle for “The Oyster Line,” a daily marine forecast serving working oyster farms. Open with two soft buoy-bell notes and a subtle rolling-water sound, then introduce an upbeat coastal folk groove with muted mandolin, upright bass, brushed snare, and a clean sonar-like chime. A confident, friendly adult woman says rhythmically: “The Oyster Line — tides, wind, and water for the working coast.” End on a memorable three-note musical logo with a crisp, broadcast-safe finish. Warm, practical, contemporary, and trustworthy; no exaggerated pirate or nautical clichés. Clear speech centered above the music, professional radio mix, no long fade-out.",
"speech": {
"volume": 10,
"speed": 8,
"pitch": 0
}
}
]'runware run bytedance:seed-audio@1.0 \
positivePrompt="Create a polished 10-second radio jingle for “The Oyster Line,” a daily marine forecast serving working oyster farms. Open with two soft buoy-bell notes and a subtle rolling-water sound, then introduce an upbeat coastal folk groove with muted mandolin, upright bass, brushed snare, and a clean sonar-like chime. A confident, friendly adult woman says rhythmically: “The Oyster Line — tides, wind, and water for the working coast.” End on a memorable three-note musical logo with a crisp, broadcast-safe finish. Warm, practical, contemporary, and trustworthy; no exaggerated pirate or nautical clichés. Clear speech centered above the music, professional radio mix, no long fade-out." \
speech.volume=10 \
speech.speed=8 \
speech.pitch=0{
"taskType": "audioInference",
"taskUUID": "3c569a98-0158-4828-af18-43495bd51734",
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a polished 10-second radio jingle for “The Oyster Line,” a daily marine forecast serving working oyster farms. Open with two soft buoy-bell notes and a subtle rolling-water sound, then introduce an upbeat coastal folk groove with muted mandolin, upright bass, brushed snare, and a clean sonar-like chime. A confident, friendly adult woman says rhythmically: “The Oyster Line — tides, wind, and water for the working coast.” End on a memorable three-note musical logo with a crisp, broadcast-safe finish. Warm, practical, contemporary, and trustworthy; no exaggerated pirate or nautical clichés. Clear speech centered above the music, professional radio mix, no long fade-out.",
"speech": {
"volume": 10,
"speed": 8,
"pitch": 0
}
}{
"taskType": "audioInference",
"taskUUID": "3c569a98-0158-4828-af18-43495bd51734",
"audioUUID": "80c974ba-c986-475c-b3af-2273b7cd5c63",
"audioURL": "https://am.runware.ai/audio/os/a10dlim3/ws/5/ai/80c974ba-c986-475c-b3af-2273b7cd5c63.mp3",
"cost": 0.026333
}Rare Book Forensics Podcast Opener
0:00
import { createClient } from '@runware/sdk'
const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()
const [result] = await client.run({
model: 'bytedance:seed-audio@1.0',
positivePrompt: 'Create a finished 35-second documentary podcast cold open. Use one close-miked host with a warm, grounded adult voice and precise neutral-English delivery; curious and quietly urgent, never theatrical. Begin with hushed archive room tone, a cotton-gloved hand opening an old hardcover, and one dry page turn. Introduce a restrained low bowed-glass texture beneath the narration, then let it resolve cleanly under the show title. Keep effects subtle, realistic, and safely behind the voice; no upbeat jingle or audience sounds. Broadcast-quality mix. Read exactly: "This book has spent a century sitting quietly on a library shelf. Its cover is a brilliant green—the height of fashion in the eighteen-fifties. But under an X-ray fluorescence scanner, that color tells a more dangerous story: arsenic. Today, conservators track the poison hidden in plain sight, and decide which books are safe to touch. This is The Quiet Evidence, a podcast about the science inside collections."',
speech: {
volume: 8,
speed: -4,
pitch: -1
}
})import asyncio
import os
from runware import Runware
async def main():
async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
results = await client.run({
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a finished 35-second documentary podcast cold open. Use one close-miked host with a warm, grounded adult voice and precise neutral-English delivery; curious and quietly urgent, never theatrical. Begin with hushed archive room tone, a cotton-gloved hand opening an old hardcover, and one dry page turn. Introduce a restrained low bowed-glass texture beneath the narration, then let it resolve cleanly under the show title. Keep effects subtle, realistic, and safely behind the voice; no upbeat jingle or audience sounds. Broadcast-quality mix. Read exactly: \"This book has spent a century sitting quietly on a library shelf. Its cover is a brilliant green—the height of fashion in the eighteen-fifties. But under an X-ray fluorescence scanner, that color tells a more dangerous story: arsenic. Today, conservators track the poison hidden in plain sight, and decide which books are safe to touch. This is The Quiet Evidence, a podcast about the science inside collections.\"",
"speech": {
"volume": 8,
"speed": -4,
"pitch": -1
}
})
asyncio.run(main())curl https://api.runware.ai/v1 \
-H "Authorization: Bearer $RUNWARE_API_KEY" \
-H "Content-Type: application/json" \
-d '[
{
"taskType": "audioInference",
"taskUUID": "ba5a1f6e-8299-4e6f-bcb5-92feebf3087e",
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a finished 35-second documentary podcast cold open. Use one close-miked host with a warm, grounded adult voice and precise neutral-English delivery; curious and quietly urgent, never theatrical. Begin with hushed archive room tone, a cotton-gloved hand opening an old hardcover, and one dry page turn. Introduce a restrained low bowed-glass texture beneath the narration, then let it resolve cleanly under the show title. Keep effects subtle, realistic, and safely behind the voice; no upbeat jingle or audience sounds. Broadcast-quality mix. Read exactly: \"This book has spent a century sitting quietly on a library shelf. Its cover is a brilliant green—the height of fashion in the eighteen-fifties. But under an X-ray fluorescence scanner, that color tells a more dangerous story: arsenic. Today, conservators track the poison hidden in plain sight, and decide which books are safe to touch. This is The Quiet Evidence, a podcast about the science inside collections.\"",
"speech": {
"volume": 8,
"speed": -4,
"pitch": -1
}
}
]'runware run bytedance:seed-audio@1.0 \
positivePrompt="Create a finished 35-second documentary podcast cold open. Use one close-miked host with a warm, grounded adult voice and precise neutral-English delivery; curious and quietly urgent, never theatrical. Begin with hushed archive room tone, a cotton-gloved hand opening an old hardcover, and one dry page turn. Introduce a restrained low bowed-glass texture beneath the narration, then let it resolve cleanly under the show title. Keep effects subtle, realistic, and safely behind the voice; no upbeat jingle or audience sounds. Broadcast-quality mix. Read exactly: \"This book has spent a century sitting quietly on a library shelf. Its cover is a brilliant green—the height of fashion in the eighteen-fifties. But under an X-ray fluorescence scanner, that color tells a more dangerous story: arsenic. Today, conservators track the poison hidden in plain sight, and decide which books are safe to touch. This is The Quiet Evidence, a podcast about the science inside collections.\"" \
speech.volume=8 \
speech.speed=-4 \
speech.pitch=-1{
"taskType": "audioInference",
"taskUUID": "ba5a1f6e-8299-4e6f-bcb5-92feebf3087e",
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a finished 35-second documentary podcast cold open. Use one close-miked host with a warm, grounded adult voice and precise neutral-English delivery; curious and quietly urgent, never theatrical. Begin with hushed archive room tone, a cotton-gloved hand opening an old hardcover, and one dry page turn. Introduce a restrained low bowed-glass texture beneath the narration, then let it resolve cleanly under the show title. Keep effects subtle, realistic, and safely behind the voice; no upbeat jingle or audience sounds. Broadcast-quality mix. Read exactly: \"This book has spent a century sitting quietly on a library shelf. Its cover is a brilliant green—the height of fashion in the eighteen-fifties. But under an X-ray fluorescence scanner, that color tells a more dangerous story: arsenic. Today, conservators track the poison hidden in plain sight, and decide which books are safe to touch. This is The Quiet Evidence, a podcast about the science inside collections.\"",
"speech": {
"volume": 8,
"speed": -4,
"pitch": -1
}
}{
"taskType": "audioInference",
"taskUUID": "ba5a1f6e-8299-4e6f-bcb5-92feebf3087e",
"audioUUID": "95647908-1f2b-4cdf-97b4-581bece7d5bf",
"audioURL": "https://am.runware.ai/audio/os/a10dlim3/ws/5/ai/95647908-1f2b-4cdf-97b4-581bece7d5bf.mp3",
"cost": 0.092167
}Semiconductor Cleanroom Gowning Safety Jingle
0:00
import { createClient } from '@runware/sdk'
const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()
const [result] = await client.run({
model: 'bytedance:seed-audio@1.0',
positivePrompt: 'Create a finished 10-second cleanroom safety jingle for a semiconductor fabrication facility. Precise, polished electro-percussion with a bright three-note glass-mallet motif, subtle filtered ventilation ambience, soft relay clicks, and a reassuring high-tech mood. A clear, confident female voice rhythmically sings: “Hood. Suit. Boots. Gloves. Check every seal.” Keep every word intelligible and evenly spaced. End with the three-note motif and a crisp, resolved sonic button. Modern industrial training audio, not childish, no sirens, no harsh alarms, no crowd noise.',
speech: {
volume: 8,
speed: 6,
pitch: 2
}
})import asyncio
import os
from runware import Runware
async def main():
async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
results = await client.run({
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a finished 10-second cleanroom safety jingle for a semiconductor fabrication facility. Precise, polished electro-percussion with a bright three-note glass-mallet motif, subtle filtered ventilation ambience, soft relay clicks, and a reassuring high-tech mood. A clear, confident female voice rhythmically sings: “Hood. Suit. Boots. Gloves. Check every seal.” Keep every word intelligible and evenly spaced. End with the three-note motif and a crisp, resolved sonic button. Modern industrial training audio, not childish, no sirens, no harsh alarms, no crowd noise.",
"speech": {
"volume": 8,
"speed": 6,
"pitch": 2
}
})
asyncio.run(main())curl https://api.runware.ai/v1 \
-H "Authorization: Bearer $RUNWARE_API_KEY" \
-H "Content-Type: application/json" \
-d '[
{
"taskType": "audioInference",
"taskUUID": "6bac8445-193c-448f-9a49-11dcabf820da",
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a finished 10-second cleanroom safety jingle for a semiconductor fabrication facility. Precise, polished electro-percussion with a bright three-note glass-mallet motif, subtle filtered ventilation ambience, soft relay clicks, and a reassuring high-tech mood. A clear, confident female voice rhythmically sings: “Hood. Suit. Boots. Gloves. Check every seal.” Keep every word intelligible and evenly spaced. End with the three-note motif and a crisp, resolved sonic button. Modern industrial training audio, not childish, no sirens, no harsh alarms, no crowd noise.",
"speech": {
"volume": 8,
"speed": 6,
"pitch": 2
}
}
]'runware run bytedance:seed-audio@1.0 \
positivePrompt="Create a finished 10-second cleanroom safety jingle for a semiconductor fabrication facility. Precise, polished electro-percussion with a bright three-note glass-mallet motif, subtle filtered ventilation ambience, soft relay clicks, and a reassuring high-tech mood. A clear, confident female voice rhythmically sings: “Hood. Suit. Boots. Gloves. Check every seal.” Keep every word intelligible and evenly spaced. End with the three-note motif and a crisp, resolved sonic button. Modern industrial training audio, not childish, no sirens, no harsh alarms, no crowd noise." \
speech.volume=8 \
speech.speed=6 \
speech.pitch=2{
"taskType": "audioInference",
"taskUUID": "6bac8445-193c-448f-9a49-11dcabf820da",
"model": "bytedance:seed-audio@1.0",
"positivePrompt": "Create a finished 10-second cleanroom safety jingle for a semiconductor fabrication facility. Precise, polished electro-percussion with a bright three-note glass-mallet motif, subtle filtered ventilation ambience, soft relay clicks, and a reassuring high-tech mood. A clear, confident female voice rhythmically sings: “Hood. Suit. Boots. Gloves. Check every seal.” Keep every word intelligible and evenly spaced. End with the three-note motif and a crisp, resolved sonic button. Modern industrial training audio, not childish, no sirens, no harsh alarms, no crowd noise.",
"speech": {
"volume": 8,
"speed": 6,
"pitch": 2
}
}{
"taskType": "audioInference",
"taskUUID": "6bac8445-193c-448f-9a49-11dcabf820da",
"audioUUID": "86352678-a978-4f81-b7c0-e43f71eefcf4",
"audioURL": "https://am.runware.ai/audio/os/a03d21/ws/5/ai/86352678-a978-4f81-b7c0-e43f71eefcf4.mp3",
"cost": 0.026333
}