Text to speech for YouTube videos, explainers and reels

Commercial usage rights come with every paid plan, from Starter at ₹499 a month. That covers a monetised channel, a client video and a sponsored reel. The Free plan is for trying voices: about 10 minutes of audio and your first 5 downloads.Commercial usage rights come with every paid plan, from Starter at $5 a month. That covers a monetised channel, a client video and a sponsored reel. The Free plan is for trying voices: about 10 minutes of audio and your first 5 downloads.

Platform rules are separate from the licence. YouTube asks creators to disclose realistic synthetic content, and its monetisation rules are about whether the video is original and useful to viewers, not about who or what read the script. Check the current creator policy before you upload, and use the disclosure setting where it applies.

The clip above is Nathan, a British English voice, reading an explainer opening, with the time at the end read the way a person says it.

Retention drops when the narration sounds flat, so this is the question that matters. In Chariot's own evaluation on English and Hindi, naturalness (NISQA) scored 4.91 in English and 4.97 in Hindi, tied for first in English and first in Hindi among the systems tested.

The voices carry mood, not just words. The clip below is Darshan, an Indian English voice, deadpan about an exam result. There are whispers, delight and plain narration in the same voice set.

There are more than 50 voices: 12 native Hindi, 8 British English, 6 American English, 5 Indian-accent voices (Assamese, Nepali, Sindhi) and 2 native voices for each of Urdu, Bengali, Marathi, Telugu, Tamil, Gujarati, Kannada, Malayalam, Odia and Punjabi. Speed is adjustable from 0.5x to 1.5x, which is how you make a 40 second read fit a 30 second short.

Ugh, I got a B+ on my exam.

PlayPauseLoadingBufferingReplayRetry

0:02

Darshan · Hindi voice

Explainers and listicles are full of things that trip up a voice: ₹1,499, 06:45, 2 GB, Dr., a URL in the call to action. Chariot reads written text the way a person says it, so you paste the script as written and the numbers, dates, times, currency, units, abbreviations and URLs come out as speech. You do not need to rewrite "₹1,499" as words or spell out the domain.

In the same evaluation, the model read hard written text (dates, currencies, URLs, units) correctly 94.0% of the time, first of nine systems.

The clip below is Alan reading a flight announcement: a flight number, a time and a gate, each read the way a person would.

Flight three zero two departs at six forty-five from gate twelve.

PlayPauseLoadingBufferingReplayRetry

0:05

Alan · American English

Most Indian creators write scripts in Hinglish, and many type Hindi in Roman letters. Chariot reads both. A sentence that moves between Hindi and English is read as one sentence, and "aapka order confirm ho gaya hai" typed in English letters is read as Hindi. The clip below is Vikram reading a Roman-script Hindi line.

For a regional version, keep the script and change the language. Chariot reads 15+ languages, and every voice can read every one of them, so your channel's voice can read the Tamil version as well as the Hindi one. Each Indian language also has its own native voices if you want a different sound per channel.

In Chariot's own Hindi evaluation, the model had the lowest Hindi character error rate (7.73%) of the seven systems tested, which is the number that tells you how often a Hindi word comes out wrong.

Namaste! Aapka order confirm ho gaya hai aur kal shaam 5 baje tak deliver ho jayega.

PlayPauseLoadingBufferingReplayRetry

0:05

Vikram · Hindi voice

Record or upload 10 to 60 seconds of yourself (about 30 seconds works best) and the cloned voice is ready in about 10 to 20 seconds. It speaks every language Chariot speaks, so a clone made from your Hindi recording can read your English video.

Cloned voices come with paid plans: 3 on Starter, 10 on Creator, 50 on Startup and 500 on Scale. You may only clone voices you have the right to use, so your own voice is fine, and a collaborator's voice needs their consent.Cloned voices come with paid plans: 3 on Starter, 50 on Startup and 500 on Scale. You may only clone voices you have the right to use, so your own voice is fine, and a collaborator's voice needs their consent.

Credits count generated audio: about 1,000 credits per minute. A 10 minute video is about 10,000 credits. A 45 second short is about 750.

The Free plan gives 10,000 credits a month, about 10 minutes, every voice, no card, and lets you download your first 5 clips. Starter at ₹499 a month gives 100,000 credits, about 100 minutes, with unlimited downloads and commercial usage rights: ten 10 minute videos or more than a hundred shorts. Creator at ₹999 gives 250,000 credits and 10 cloned voices. Annual billing gives 2 months free.The Free plan gives 10,000 credits a month, about 10 minutes, every voice, no card, and lets you download your first 5 clips. Starter at $5 a month gives 100,000 credits, about 100 minutes, with unlimited downloads and commercial usage rights: ten 10 minute videos or more than a hundred shorts. Annual billing gives 2 months free.

In the web app at platform.chariot.in you paste the script, pick a voice and download MP3 or WAV. For a full-length script with video, use Studio. If you generate narration from code, POST https://api.chariot.in/v1/tts returns a WAV file and takes up to 1,000 characters per request, so send a script a paragraph at a time.