By ShinobiTools Team · Last updated: August 2026
Paste your text, pick a voice and hear it straight away. 5,000 characters per go, no daily cap, no account — and the MP3 download is included, not sold separately. Whether you call it text to speech or text to voice, the process is the same: paste, pick a voice, press play.
Most "free" text to speech generators meter you: an account first, then a monthly allowance of credits, and the download button behind a paywall. The best-known one gives you 10,000 credits a month and asks you to sign in before you hear a word.
This page works the other way round. Type or paste up to 5,000 characters, press one button and listen. No account, no credit counter, no daily cap — when a text is longer, split it and run the parts back to back. The speech is generated on our own hardware, which is why we don't have to meter it.
It is also the mirror of what this site is known for: audio to text turns speech into writing, this turns writing into speech.
A voice you haven't heard is a voice you can't judge, so the result arrives as a player first: press play, and if it sounds right, the MP3 download sits next to it. No preview-with-a-catch — the file you hear is the file you get.
There are five voices: three American (two female, one male) and two British (one female, one male). They read naturally — sentences flow, questions rise, abbreviations are spoken out — rather than the flat robot voice older generators produce.
A paragraph takes a few seconds; the full 5,000 characters comes back as roughly six minutes of audio in about a minute of processing.
The voices are English — American and British — for now. Other languages follow when they pass the same quality bar; a language done badly is worse than a language missing.
It reads, it does not clone. You cannot upload someone's voice and make it say things — the consent problems there are real, and we would rather skip the feature than pretend they aren't.
It reads prose best. Tables, code and heavy math notation come out the way they would if a person read them aloud cold — plain sentences give the cleanest result.
The text is processed on our own hardware in the Netherlands, not sent to a third-party API. It is not stored, not logged and not used to train anything.
The generated audio gets a private link and is deleted automatically after 45 minutes — download the MP3 and it is yours; leave it, and it is gone. See privacy for the full picture.
Yes. No account, no credit meter, no daily cap and no watermark-style audio tags. 5,000 characters per go, as many goes as you like. Ads around the tool pay for the hardware.
Yes. The MP3 is yours to keep and use, including in monetised videos and commercial projects. No attribution needed.
English, in American and British voices, for now. More languages will follow once they pass the same quality bar as these five voices.
5,000 characters per go — roughly six minutes of audio. There is no daily cap, so split a longer text into parts and run them one after another.
No. It is processed on our own hardware, the audio is deleted automatically after 45 minutes, and neither the text nor the audio is used for training.
No, deliberately. You pick from five fixed voices; uploading someone's voice to imitate them is a consent minefield we stay out of.
Unusual names and invented words are read phonetically, the way a person seeing them for the first time would. Ordinary prose reads naturally.
Not publicly. If you have a project that needs one, mail [email protected] and tell us what you're building.