Milton
Male · Young · Multilingual
Weather report
Style: Read in Japanese like a TV weather announcer: even, precise, unhurried.
気象庁によりますと、今週末は全国的に晴れる見込みです。
Japanese AI text to speech
with natural voices
Japanese text to speech with narrators you direct in plain words, for explainer videos, staff training, phone lines, and audiobooks. When a kanji is read the wrong way, write it in kana, and set casual or polite speech in the style line. Free plan: 1,000 words, no credit card.
Enbee V2 narrators
21 multilingual narrators
Paste Japanese text, kanji and kana as written
Use full-width punctuation as you normally would. If a name or word comes out wrong, rewrite it in hiragana and play it again.
Script language
Choose a narrator
61 / 500
Script
33 / 250 charactersTags to try: [laughs] [whispers] [sighs] [angry] [excited] [sarcastic]
Kanji readings, counters, and keigo
Most mistakes in Japanese narration come from a kanji with more than one reading or a counter that changes its sound. Here is how to catch and fix each one.
One kanji, two readings
辛い（からい／つらい）
辛い can mean spicy (からい) or painful (つらい), and 今日 is usually きょう but こんにち in 今日は. Context usually settles it. When it does not, or when a personal name reads wrong, write that word in hiragana or katakana.
Counters change the sound
1本 いっぽん、3本 さんぼん
The number and the counter affect each other: 1本 is いっぽん, 3本 is さんぼん, 4月 is しがつ, and the first of the month is ついたち. Listen to every counted item in the preview, and put the reading in kana if one slips.
Keigo level goes in the style line
待って／お待ちください／少々お待ちくださいませ
Casual, polite, and very polite versions of "please wait". The words carry most of the politeness, so write the level you need, then describe the delivery in the direction, for example "polite keigo, calm". Support and IVR scripts normally use the polite forms.
Era years and full dates
令和8年 / 2026年4月1日
令和8年 is 2026 and is read れいわはちねん. 2026年4月1日 is read にせんにじゅうろくねん しがつ ついたち. Keep the form your audience expects, era year for official notices and Western year for most videos.
Katakana loanwords and natural delivery
コンピューター、アップロード
Loanwords written in katakana are read the Japanese way. Explainer channels have long used flat synthetic voices; a narrator you can direct, with pace and mood in the style line, gives a more natural read for the same script.
iPhone, PDF, and 3秒 in a Japanese sentence
Japanese product scripts drop in Latin-letter brand names and abbreviations. Leave them in Latin letters and the narrator reads them the way Japanese speakers say them.
An app feature line with a brand name, an abbreviation, and a count
Style: Speak in Japanese, clear and upbeat, like an app demo video
新しい iPhone アプリで PDF を 3秒で 音声に変換できます。
Each part of the app line
- Kanji and kana
- 新しい、音声に変換できます
- Latin brand names
- iPhone
- Abbreviations
- Number with a counter
- 3秒で
Prices, dates, era years, and times in Japanese
Write numbers in Arabic digits with their counters. The reading changes with the counter, as these rows show.
| You write | The narrator says | Type |
|---|---|---|
| 1,980円 | せんきゅうひゃくはちじゅうえん | Price in yen |
| 2026年4月1日 | にせんにじゅうろくねん しがつ ついたち | Irregular 4月 and 1日 |
| 令和8年 | れいわはちねん | Era year for 2026 |
| 午後3時15分 | ごご さんじ じゅうごふん | 分 is ふん or ぷん by number |
Four Japanese narrators at four levels of formality
A weather report, a YouTube opener, an app demo, and a keigo phone greeting. Compare how the same narrator set handles plain, friendly, and very polite Japanese.
Russell
Male · Middle-Aged · Multilingual
YouTube explainer opener
Style: Speak in Japanese with a friendly, curious tone, like a tech channel host.
今日は、知らないと損するスマホの設定を3つ紹介します。
Ivy
Female · Young · Multilingual
App product demo
Style: Speak in Japanese, clear and confident, walking through a feature.
このアプリなら、PDFをアップロードするだけで音声に変換できます。
Lenora
Female · Middle-Aged · Multilingual
Phone greeting in keigo
Style: Speak in polite Japanese keigo, gentle and composed, like a support line operator.
お電話ありがとうございます。担当者におつなぎしますので、少々お待ちください。
New-hire training audio, one block per part
A security module for new employees: a greeting, the topic, and a closing line. HR can rewrite a rule next quarter and regenerate just that block.
01Milton
Style: Speak in polite Japanese, warm but businesslike, welcoming new staff.
皆さん、こんにちは。新入社員研修へようこそ。
02Lenora
Style: Read in Japanese at a measured pace, as a trainer introducing the module.
このモジュールでは、社内のセキュリティルールを3つ説明します。
03Milton
Style: Speak in Japanese with a courteous close, slightly slower.
以上で第1回は終了です。次回もよろしくお願いします。
Policy updated, module mostly unchanged
Edit the block that names the rules and regenerate only that one. The greeting and the closing line keep the audio staff already know.
Formality and mood in writing
Describe the register in the style line, such as "polite, calm" or "casual, upbeat", and put [softly] or [excited] in front of a phrase that needs its own mood.
Block-by-block onboarding narrationCompany and product names
Enbee V1 voices can save the reading of a company or product name once, and give you speed and pause controls for every module after.
Three tools, depending on what you are narrating
Video and training scripts belong in the Studio, full books in the Audiobook Creator, and recurring personas in voice cloning.
Studio for explainers and training
Lay a script out as blocks, each with a narrator and a formality level, and reorder them by dragging. Change a sentence and regenerate only its block.
Lay out a 研修 or 解説 script in blocksAudiobook Creator, with kana fixes per paragraph
Upload EPUB, PDF, or DOCX and the book is divided into chapters. When a kanji reading is off, correct the paragraph in kana and regenerate that paragraph alone.
Produce a Japanese audiobook from EPUBOne persona voice across every upload
Clone a voice you own, or have permission to use, from a short sample, and keep one recognisable voice across a channel or a character.
Create a persona voice from a sampleStart on the free plan with 1,000 words and no credit card; downloads from it are watermarked. Paid plans add words, remove the watermark, and include commercial use for client work and published titles. Narration Box pricing in detail
Japanese voiceovers by job
Japanese audio is used across Japan and by Japanese-speaking communities in Brazil and the US, on YouTube, Niconico, audiobook apps, VTuber streams, and company learning systems.
History and culture explainers
Long-form YouTube explainers read by a narrator you can direct, instead of the flat voice the genre is known for.
今回は、日本の城の歴史をわかりやすく解説していきます。
Japanese explainer channel narrationCompliance and HR modules
Short courses for staff, with polite delivery and a block per topic so updates stay small.
この研修は約10分で終わります。
Company training narrated in JapaneseMystery, romance, and literary novels
Import a manuscript, let the chapters split, and fix individual paragraphs when a reading needs kana.
その朝、駅のホームには誰もいなかった。
Audio editions of Japanese fictionPhone menus and hold messages
IVR prompts in keigo that match the tone of your live operators.
ご用件の番号を押してください。
Keigo IVR promptsOnline store promotions
First-purchase offers and seasonal sales for e-commerce videos, in a few takes to pick from.
今なら初回購入で20%オフ。
初回購入 offers voiced for online shopsGame and VTuber characters
Character lines with inline mood tags, or a cloned voice you own the rights to for a recurring persona.
[excited] やった！ついに見つけたよ！
Cloning a character voice you ownGoing beyond Japan with the same script
Teams localising a Japanese video or course often need Korean, Chinese, English, or Portuguese versions as well. Or browse all 80+ text to speech languages.
Guides for explainers, demos, and audiobooks
Natural-sounding narration, multilingual product demos, and directing emotion in a voiceover.
Jan 20, 2026
How to Make an Audiobook Sound Natural With AI
Read article
Jul 16, 2025
AI Voice tool to make multilingual product demos
Read article
Feb 28, 2026
How to add emotion without sounding fake (style prompts + examples)
Read article
Nov 27, 2025
How to create multilingual onboarding tutorials: 2026
Read article
Customer testimonials
See why creators, publishers, and teams rely on Narration Box for natural-sounding voice content.
“Quality of voices and ease of use made Narration Box the perfect choice for my fiction podcast The Program. It's the only voice synthesis service that knew the difference between 'live frugally' and 'live broadcast' and that could pronounce Mar-a-Lago.”
Ivan
Creator, The program audio series
“We've been using Narration Box for a while now, and are grateful to the team behind it! It allows us to create audio versions of our books and articles in many languages, within seconds. The generated audio sounds very natural and is almost indistinguishable from a real narrator. What an amazing service for creators, and it keeps evolving! I highly recommend it.”
Reto Stuber
CEO, Stuber Media
“The user-interface is great. Nice and simple without too many bells and whistles. The array of voices are also good, though some are definitely better than others. With that said, you can adjust the speaking pace for each voice, which does improve the flow, making voices seem less artificial. Also, the free plan is generous.”
Tom W.
Filmmaker
Review on G2
“Narration Box has allowed me to get away from having to set up my own recording studio. Previously, I had an idea of hiring a human narrator but there was so much hassle as well as costs. The voice choices are pretty good, especially the enhanced voices. I love the options available to further personalize the AI narrator's speech within the text. Narration Box saved me so much time for several small projects. I actually made audio clips whilst writing another book.”
Aammton Alias
M Content Creations
FAQs
Frequently asked questions
Readings, keigo, era years, accents, and licensing for Japanese narration.
Japanese text to speech
Can I test Japanese narration before I subscribe?
Yes. The free plan covers 1,000 words with no credit card. Free downloads carry a watermark, so treat them as tests and move to a paid plan before publishing.
Which voices can narrate Japanese, male and female?
Narration Box offers 1,500+ AI voices across 80+ languages, and that includes Enbee V1 voices made for Japanese. The clips on this page use our 21 Enbee V2 narrators (12 female, 9 male), each of which reads Japanese and follows written direction on pace, mood, and politeness.
A kanji or a name is read wrong. How do I fix it?
Rewrite that word in hiragana or katakana, for example からい instead of 辛い when you mean spicy, and regenerate the block. For a name used across many projects, an Enbee V1 voice can save the reading.
How do I get keigo for a support line?
Write the script in the polite forms you want, such as 少々お待ちください, and describe the delivery in the style line, for example "polite keigo, calm and clear". The style line shapes the delivery, while the politeness itself lives in the words you write.
Are era years like 令和8年 read correctly?
Yes, 令和8年 is read れいわはちねん. Western dates such as 2026年4月1日 are read with the irregular しがつ and ついたち. Preview any date with an unusual counter.
Can I get a Kansai accent or a non-native Japanese accent?
Ask for a Kansai dialect feel or a foreign-learner accent in the style line. How far a narrator goes varies, so listen to a short block first, then pick whoever sounds most convincing.
Can I make a Japanese audiobook with it?
Yes, with the Audiobook Creator, which imports EPUB, PDF, or DOCX and splits chapters for you. Check your retailer or distributor rules for AI narration, since acceptance is up to them.
Can my VTuber persona speak in my cloned voice?
Yes, if it is your voice or you have permission to use it. Create the clone from a short sample, then listen to a few Japanese lines first, since accent carry-over from another language varies.
Can the audio go into a client course or a paid app?
Yes, on a paid plan, under the subscription terms. Pick MP3 for video editors and LMS uploads, or WAV, Opus, FLAC, or OGG, delivered as one track or split into a file for each block.
Get Started with Narration Box Today
Choose from our flexible pricing plans designed for creators of all sizes. Start your free trial and experience the power of AI voice generation.