Japanese text to speech
Japanese blends three writing systems inside a single sentence, and the most demanding of them is ambiguous by design. Most characters carry at least two readings — one inherited from Chinese, one native — and which applies depends entirely on the compound the character sits within. A personal name may take a reading found nowhere else, chosen by the family that bears it.
Try the live product
Type, choose a voice, then press play.
Type a script, pick a voice, and hear it read back.
Free to try — no sign-up.
60/500Pitch accent then separates words sharing every one of their sounds. This is neither stress in the English sense nor tone in the Chinese one; it is where the pitch drops, and standard Tokyo speech relies on it to keep apart vocabulary that would otherwise be indistinguishable to the ear.
Because sentences are written without spaces and move freely between scripts, segmentation and reading selection have to be solved together rather than in sequence. That combination makes Japanese one of the harder languages to narrate convincingly, and one of the most obvious when it has been done badly.
No signup needed — the demo speaks up to 500 characters.
Regional delivery
- Japanese
Every voice speaks every supported language, so one narrator can carry your whole catalogue across markets.
Japanese production details
Readings chosen by compound
Which pronunciation a character takes follows the compound around it instead of defaulting to one fixed reading per glyph.
Pitch accent carried
Where the pitch falls is reproduced as lexical information, since standard speech uses it to separate otherwise identical words.
Three scripts, one sentence
Characters, both kana sets and embedded Latin text are read together without the delivery stumbling at each transition.
Irregular name readings
Personal and place names taking an unpredictable reading can be saved as rules so later scripts never revert to the common form.
56 studio voices (30 premium, paid plans) · 20 output languages · SRT captions and bulk mode on every plan
Free to start
Use the free plan first, then move to a monthly plan when your production volume grows.
25,000 characters
Free plan monthly allowance
$7/month
About 3 hours on the Creator plan
Questions, answered
How does it decide between competing character readings?
The compound a character appears in drives the decision, rather than a single fixed default per character. This is why context matters considerably more here than raw character coverage does.
Can I force one specific reading for a name?
Yes, and for personal names you frequently should. Family names in particular take readings no general rule predicts, so one saved rule per name is worth the small amount of setup.
Is pitch accent actually reproduced?
Yes. Pitch accent separates words that are otherwise identical in sound, so a flat reading would collapse distinctions listeners depend on to follow a sentence at normal speed.
Does embedded Latin text inside Japanese work?
Yes. Latin-script words and figures are read where they sit, which matches the way technical and business Japanese is normally written in practice.
Should I annotate difficult words with reading aids?
Usually unnecessary. Where a word is genuinely ambiguous or a name is irregular, a saved pronunciation rule is more reliable and more durable than annotating the source text each time.