Text to speech (TTS)
Turn any text into natural, expressive speech. Describe the tone in plain language to shape the delivery.
Text to speech · Voice cloning · Multilingual · Speed/pitch control · MP3 / WAV export
Seed Audio 1.0 — AI Voice Generation
Generate natural, expressive AI speech from text or reference audio. Seed Audio 1.0 supports 20 preset voices, voice cloning, and speed/pitch control.
Natural speech from text
Type your script and get expressive AI speech in seconds.
Voice cloning
Reproduce a voice from up to 3 reference clips (30s each).
Multilingual
Supports many languages, including Japanese.
Loading generator...
Seed Audio 1.0 voice samples
Examples generated with Seed Audio 1.0 — listen to the real voices right here.
Podcast
「今日のテーマは、AI 音声の活用法です。台本を貼り付けるだけで、自然な会話のような読み上げが作れます。」
0:00 / 0:06
Voice cloning
「ブランドの声を、すべてのコンテンツで統一できます。一度設定すれば、同じトーンで何度でも生成できます。」
0:00 / 0:08
Multilingual
「Seed Audio は日本語と English など、複数の言語に対応しています。」
0:00 / 0:05
Natural delivery
「わぁ、すごい！こんなに自然な音声が、本当に数秒でできるなんて驚きです。」
0:00 / 0:06
Features
What Seed Audio 1.0 can do
From text to natural speech. Text-to-speech, voice cloning and multilingual output in one generator, with flexible controls for any use case.
Voice cloning
Reproduce a voice from reference audio (up to 3 clips, 30s each). Keep one brand voice across all content.
20 preset voices
Pick from narration, character or news voices to match your tone and use case.
Multilingual
English, Chinese, Japanese, Spanish, Indonesian and Portuguese, including mixed-language text.
Speed, pitch & volume
Fine-tune speed (0.5–2.0x), pitch (±12 semitones) and volume.
MP3 / WAV export
Export as MP3, WAV, PCM or OGG Opus, up to 48 kHz.
Use cases
What you can build with Seed Audio
For any workflow that needs a voice. Creators, developers and teams ship audio fast — without a recording studio.
Video voiceovers
Narrate YouTube videos, ads and explainers with a consistent voice. Regenerate a line in seconds when the script changes.
Podcasts & audiobooks
Turn long scripts into clear narration that keeps tone and pacing steady across chapters.
News & announcements
Pick a steady voice and speed for a clear, broadcast-style read.
Voice cloning
Build a reusable brand voice from a short sample and keep it consistent across videos, courses and social.
Multilingual content
Speak 6 languages, so videos and courses for global audiences use the same workflow.
Apps & games
Add speech to assistants, menus and game dialogue, and more.
How it works
Generate voice in 3 steps
Enter text and generate AI speech right away — no complex setup.
- 01
Enter your text
Type the script you want to voice. Add a mood like “calm tone” to shape the delivery.
- 02
Choose voice & settings
Pick from 20 preset voices and adjust speed, pitch and volume. You can also clone a voice.
- 03
Generate & export
Generate in seconds, preview, and export as MP3 / WAV. Regenerate any line to fine-tune.
Answers
Frequently asked questions
Answers about Seed Audio 1.0.
An audio generation model by ByteDance. It generates natural speech from text or reference audio, with text-to-speech (TTS) and voice cloning, usable right in your browser here.
Yes. It supports English, Chinese, Japanese, Spanish, Indonesian and Portuguese, generating natural speech from Japanese text.
Yes — from reference audio (up to 3 clips, 30 seconds each), referenced in the prompt as @Audio1–@Audio3.
New accounts get free credits to try your first generations, then you spend credits based on usage.
Please review your intended use and the applicable terms. See the pricing page and FAQ for details.
MP3, WAV, PCM or OGG Opus, with sample rates up to 48 kHz.
Generate AI voice now
Create natural speech from text with Seed Audio 1.0. Your first generation is covered by free credits.