Accessibility
Text to speech for accessibility and listening
Screen readers already exist, blind and low-vision users are typically expert with them, and a synthetic voice reading a page aloud is no substitute for a document that is structured, labelled and navigable. If your content is inaccessible, audio does not fix it — it produces an inaccessible thing that talks.
What generated audio genuinely does is make written material listenable for people who find reading it hard or impractical: readers with dyslexia, people whose literacy in a language is weaker than their fluency in it — which in India is an enormous group — anyone commuting, and anyone who simply retains more by listening.
The workflow
- 1
Fix the document before generating anything
Proper headings, alt text, meaningful link text, sane reading order. This helps screen-reader users, who are who accessibility rules are actually about, and it also produces better audio because the structure survives into it.
- 2
Edit the text for listening
Strip what only works visually: "see the table below", "click here", parentheticals three clauses deep. Spoken text needs the information in the order the ear takes it.
- 3
Generate in the reader's language, not the document's
Where a Hindi-speaking audience is reading English material, a Hindi audio version reaches people the English text never did. This is the highest-value version of this use case in India and it is mostly unexploited.
- 4
Publish audio alongside text, never instead of it
Audio is an additional route, not a replacement. Removing the text to "simplify" makes content less accessible, not more.
- 5
Keep one voice across a body of work
For a course or a publication, a consistent voice reduces the effort of listening. A cloned voice does that permanently — the same voice a year from now.
What goes wrong, and what to do about it
- Treating audio as a substitute for accessible structure.
- It is not one, and regulators do not accept it as one. Fix headings, labels and reading order first; audio sits on top of that.
- The audio reads out visual furniture — captions, footnote markers, table cells.
- Prepare a listening version of the text. It is a five-minute edit per article and it is the difference between usable and unbearable.
- Long documents become one enormous audio file.
- Generate per section and publish with clear section titles so listeners can navigate. Nobody scrubs through a fifty-minute file to find a paragraph.
- Numbers, abbreviations and acronyms come out wrong.
- Write them the way they should be spoken in the listening version. It matters more here than in entertainment content, because the listener has no text in front of them to check against.
Which plan you actually need
Roughly 800–900 characters of English is a minute of audio, so a 1,500-word article is about ten minutes. Starter's 60 minutes covers around six articles a month; a publication running daily audio versions is in Premium's 250 minutes or Premium Plus's 800.
Full limits are on the pricing page. Every plan includes the same 34 languages — plans differ on volume, voice slots and queue priority, not on capability.
Frequently asked questions
Does adding audio make my site accessible?
No. Accessibility is about structure, labelling and navigability — the things assistive technology relies on. Audio is a genuinely useful addition for readers who prefer or need to listen, and it does not substitute for accessible markup or satisfy accessibility requirements on its own.
Is this useful for readers in Indian languages?
It is arguably the strongest version of the use case. Many people are far more fluent in a language than they are literate in it, so an audio version in Hindi, Tamil or Bengali reaches an audience the written version does not — including for material originally written in English.
Can I publish generated audio versions commercially?
Yes, on a paid plan. Free-tier output is watermarked and English-only, which makes it unsuitable for a published audio programme.
Other use cases
- AI voice for YouTube videosNarrate YouTube videos in your own cloned voice or a preset one, in Hindi, Hinglish or English. Consistent across episodes, and cheap enough to re-render a line you got wrong.
- AI voice for Reels, Shorts and short-form videoTurn a 30-second script into voiceover in seconds, in Hinglish or Hindi. Built for creators publishing daily, where recording every clip is the bottleneck.
- AI voice for podcastsGenerate podcast intros, ad reads, corrections and pickup lines in your own cloned voice, without booking studio time for thirty seconds of audio.
- AI voice for audiobooksNarrate books and long-form documents in a consistent voice, with sentence-boundary chunking that keeps chapter-length text from breaking mid-clause.
Try it for accessibility
Start on the free tier with your own script. If it does not sound right, you have lost five minutes.