Turn written content into a clear read-aloud.
Articles, documentation, and help content are not always easy to read. Generate a clean spoken version so the same content reaches people who prefer or need to listen.
No credit card · Real engines · The audio is yours
Your help center is four hundred articles deep, and every one of them assumes the reader can comfortably read. Support keeps hearing from people who would rather listen. Recording a human read of a knowledge base that changes weekly was never going to happen.
What Accessibility actually needs.
A lot of written content stays written, which leaves out anyone who reads with difficulty or simply prefers to listen. Recording a human read of every article and doc page is not realistic at the volume content moves. Generating a clear spoken version means the same material can be offered as audio without a recording session for each update.
Real features, mapped to the job.
Every item here works today, or says plainly where it is still in progress.
Clear, neutral read
Kokoro gives a clean, plain read that suits articles and documentation, where clarity matters more than drama.
Keeps pace with updates
When the content changes, regenerate the audio. The flat allowance means keeping audio current does not run up a per-character bill.
Long-form in one pass
Up to 30,000 characters per generation, so a full article or doc page goes through without chopping it up.
Own and host it
Export MP3 or WAV with commercial rights and no watermark, yours to host alongside the written version.
Read-aloud: a help-center article
This article is also available as audio. Press play to listen.
To export your data, open Settings and choose Download archive. The file arrives by email within a few minutes.
If the download does not arrive, check your spam folder first, then contact support from the same page.
Kokoro, voice Michael: a steady, plain read. An audio alternative serves the spirit of accessible content, but it is not a screen reader and does not make a page WCAG-conformant on its own. We will not claim otherwise.
- ~12,000
- characters in a 2,000-word article
- ~12 min
- of listening from that same article
- 30,000
- characters per pass, no chopping a long page
How it goes, step by step.
Step 1: Paste the content
Drop the article or doc text into Text to Speech, up to 30,000 characters per pass.
Step 2: Pick a clear voice
Choose a clean, neutral voice that reads plainly for the spoken version.
Step 3: Generate and publish
Generate the audio, export it, and host it next to the written content.
What an audio version adds to accessibility, and what it does not.
It complements assistive tech, never replaces it
People who rely on screen readers already have one, tuned to their own speed and habits. An audio edition serves a different group: readers with fatigue or dyslexia, people newer to your language, anyone listening on the move. Treat generated audio as one more accessibility option on the page, not as the page's accessibility strategy.
Conformance lives in the markup
Publishing an audio edition does not make a page meet WCAG, and a vendor claiming otherwise should make you suspicious. Conformance is structure: headings, alt text, contrast, keyboard paths. The audio file is a courtesy on top, valuable precisely because nobody is pretending it is more.
Stale audio is its own failure
An outdated spoken version hands the listener yesterday's instructions for today's interface, which is worse than no audio at all. Regenerating the read when an article changes takes minutes on a flat allowance, so the spoken edition can carry the same revision date as the written one.
Start with Kokoro.
Kokoro reads cleanly and plainly and drafts fastest, which fits read-aloud of articles and documentation where clarity is the priority.
Cheapest. Clean, plain read. Ignores cues.
- Quality Elo
- 1060
- Latency
- 973 ms (measured 2026-06-10)
- Languages
- 8
- Rights
- Apache-2.0 model; commercial OK
“This article is also available as audio. Press play to listen instead of reading.”
The honest answers.
What Cantari can and cannot do for accessibility today, in plain language.
Is this a full assistive-technology replacement?
Which voice is best for read-aloud?
Can I update the audio when the article changes?
Try Cantari for accessibility.
Free to start, no credit meter. Open the studio and hear it for yourself.