Start with your scene
Write the line, describe the voice and atmosphere, then add one optional reference when your character needs a stronger anchor.
音频参考与图像参考不可同时使用。
输入 @ 搜索声音或引用已上传的音频
AI 语音
Seed Audio 生成需要先登录。登录后可使用你的 Credits 生成语音。
输入文本，并可选择音频或图像参考，两者不能同时使用；生成结果支持试听和下载。
包含: Natural AI voice · In-browser preview · MP3 and WAV download · Starter trial credits - about 30 seconds of audio · 自然的 AI 语音 · 浏览器内预览 · MP3 和 WAV 下载 · 入门试用 Credits——约 30 秒音频
描述你的恐怖语音和场景。
输入文本，并可选择音频或图像参考，两者不能同时使用；生成结果支持试听和下载。
You direct pace, emotion, character, ambience, and dramatic shifts instead of settling for one deep preset.
Describe pace, timbre, emotion, dialogue, and ambience in one Seed Audio 1.0 prompt. You move from a plain read to a designed horror performance. You can reuse a prompt template when another scene needs the same creative structure.
Add one self-owned audio reference or one non-identifiable character image, never both together. Your voice starts from a clearer character anchor. You can choose the reference type that matches the material already in your workflow.
Keep one async run in history, retry failure, preview playback, then download the result. You move from waiting to a playable file without restarting your brief. You can compare the preview with your intended scene before choosing the download action.
在一个 Seed Audio 1.0 提示词中描述节奏、音色、情绪、对白和环境氛围。 把平淡朗读变成经过设计的恐怖演绎。 当其他场景需要相同创作结构时，你可以复用提示词模板。
添加一个自有音频参考，或一张无法识别真实身份的角色图像，两者不可同时使用。 让生成的声音从更清晰的角色基准出发。 你可以根据现有创作素材，选择最合适的参考类型。
把异步任务保留在历史记录中，失败后重试，试听播放后下载结果。 无需重新填写创作要求，从等待直接进入可播放文件。 决定下载前，你可以先对照预想场景判断试听结果。
Choose a real creative job, from restrained narration to a full haunted-game exchange, and give each moment its own sound.
Turn a story passage into calm, eerie narration with a consistent self-owned voice reference. You pair one self-owned reference clip with a short creepypasta passage and leave with playable horror narration voice audio.
Create a memorable demon or monster voice for a boss encounter without recording the part. You turn one boss line into a gravelly demon voice generator clip for a cutscene or encounter prototype.
Prototype an unsettling exchange between distinct characters inside one generation. You generate a two-character haunted game dialogue clip with a childlike ghost and a guarded night watchman.
Build a cold-open voice and atmosphere that establishes the episode in seconds. You turn one episode tease into a scary podcast intro with voice, a two-second sound bed, and a closing sting.
Add an immediate emotional voice hook to a short Halloween reveal or trailer beat. You generate a Halloween voiceover that opens at peak panic and lands a named door-slam sting for the edit.
Keep each async generation in history, retry a failed run, preview the result, and download the audio you want to use. 从提示词到下载，全程掌控. 你可以用经过验证的输入方式和清晰可见的恢复步骤，测试完整的 AI 恐怖语音生成器流程，而不必相信未经证实的质量承诺。
You enter text and can add either reference audio or a reference image when the character needs an anchor.
You preview generated audio before choosing the download action.
You keep async generations in history and can retry after failure.
You receive starter credits on the free plan.
输入文本；当角色需要参考基准时，还可以添加参考音频或参考图像，二选一。
选择下载前，可以先试听生成的音频。
异步生成会保留在历史记录中，失败后可以重试。
免费套餐会提供入门 Credits。
音频参考与图像参考不可同时使用。
免费套餐包含入门 Credits。
发布生成音频前，请查阅适用于你的目标项目的最新使用条款。
Try the workflow first
Use the free plan's starter credits to test a short horror line, a character direction, and the final preview flow.
免费套餐包含入门 Credits。
使用入门 Credits 试用 Seed Audio 生成
适合高用量 Seed Audio 月度生成
Check how prompts, references, free starter credits, previews, downloads, history, and retries work before your first run.
An AI scary voice generator turns your text direction into eerie narration or character audio. You describe the voice, pace, emotion, and scene, then preview the generated result.
Describe the character, delivery, pace, ambience, and emotional turn in the prompt. You can ask for a calm whisper, a controlled growl, or a sudden panic shift.
Yes. You can direct a demon or monster character with natural-language details such as gravelly timbre, measured menace, a growled delivery, and a short ambience cue.
Yes. You can add optional reference audio or one character image to guide the result, but you cannot combine audio and image references in the same run.
Yes. You can use the free plan's starter credits to test a short script and the preview-and-download workflow.
Yes. You can play the generated audio in the result view and use the download action when the result fits your scene.
You keep the generation in history and can retry after failure instead of rebuilding the request from scratch.
You can create horror narration, demon and monster characters, haunted game dialogue, scary podcast intros, and Halloween social-video voiceovers.
Review the current usage terms for your intended project before publishing or monetizing generated audio; this page does not promise rights beyond those terms.
AI 恐怖语音生成器会根据文本指令生成诡异旁白或角色音频。你可以描述声音、节奏、情绪和场景，再试听生成结果。
在提示词中描述角色、表达方式、节奏、环境氛围和情绪转折。你可以要求冷静低语、克制低吼或突然转为恐慌。
可以。你可以用自然语言细节指导恶魔或怪物角色，例如粗粝音色、有分寸的威胁感、低吼式表达和简短的环境声音提示。
可以。你可以添加参考音频或一张角色图像来引导结果，但同一次生成不能同时使用音频和图像参考。
可以。你可以使用免费套餐的入门 Credits，测试一小段脚本以及试听和下载流程。
可以。你可以在结果页面播放生成的音频；结果符合场景需求时，再使用下载操作。
生成任务会保留在历史记录中，失败后可以重试，无需从头重新填写请求。
你可以创作恐怖旁白、恶魔和怪物角色语音、闹鬼游戏对白、恐怖播客片头，以及万圣节社媒视频配音。
发布生成音频或用于变现前，请查阅适用于你的目标项目的最新使用条款；本页面不会承诺超出这些条款的权利。
