How We Make the Stories and Narrations

Where Story Voice stories come from: public-domain originals retold by us in English and Chinese, narrated by ElevenLabs AI voices, all labeled as AI.

Last updated: 2026-11-11

Every story on Story Voice comes from a public-domain original and is retold by us, in English and in Chinese. The narration is generated by AI voices and labeled as AI-generated. Here is how we do it.

Where the stories come from

Public-domain originals only: the Brothers Grimm's Children's and Household Tales (1812–1857), Hans Christian Andersen's fairy tales (1835–1872), Aesop's fables, and fables and idiom stories from ancient Chinese books such as the Han Feizi, the Zhuangzi, the Strategies of the Warring States (Zhanguo Ce) and the Lüshi Chunqiu. Their authors died more than a hundred years ago, and the original texts are no longer protected by copyright.
Our own retellings, not existing translations: many translations are still under copyright (for example, Ye Junjian's Chinese translation of Andersen). We don't copy any existing translation; we rewrite both the Chinese and the English versions ourselves from the plots of the originals. The English retellings are our own too, written to be read aloud.
Retold for bedtime: the originals contain quite a few gory or frightening details. We cut them or make them gentler, and at the bottom of each story page, under "Where this story comes from", we give the original source and say what we changed.
No picture books: we never include copyrighted picture books, cartoons or school textbook texts.

What does the narration

All narration on the site is generated by AI voice models from ElevenLabs, not recorded by people:

Use	Model	Notes
Chinese narration	ElevenLabs Eleven v3	A gentle, motherly voice designed for the site, about 180 Chinese characters a minute
English narration	ElevenLabs Eleven Multilingual v2	A British storytelling voice, slowed to about 148 words a minute so children can read along
Reading in your voice	ElevenLabs Instant Voice Cloning + Eleven Multilingual v2	Clones only your own voice; the clone is deleted right after use
Writing a new story	DeepSeek large language model (deepseek-chat)	Writes an original story from the main character and lesson you give, then the story is checked for content

After each narration is generated, we play it back through speech recognition to check for skipped or misread sentences. Even so, AI narration may occasionally mispronounce a word; if you notice one, please let us know.

How your voice is handled

Only your own voice: before recording, you give separate consent and read aloud a random sentence shown on the page, to confirm that it is you. You cannot record family members, children or anyone else.
Deleted after use: the cloned voice is used only for this one reading. It is deleted from the voice provider as soon as the reading is finished, and it is not kept or used for anything else.
Only you can listen: your recording and the audio read with it are kept in private storage and can only be played when you are signed in to your account. You can delete your recording at any time.
Labeled as AI-generated: generated audio files carry a built-in AI-generated label.

Content safety

Every generation must pass a content check first: the site's list of blocked words, a safety judgment by a large language model and, for story requests you type yourself, Waffo's Prompt Sift check. A newly written story is checked again before it is narrated. Anything that fails is not generated, and no credits are deducted. See the Acceptable Use Policy for the full rules.

Who is behind it

Story Voice is operated by [TBD: operator], and payments are processed by Waffo as merchant of record. If you find a typo, a mispronunciation or anything not right for children, or would like to suggest a story, write to 【待定：联系邮箱】, or sign in and tell us under "Settings → Support".