Text to speech for accessibility teams
Built around alternative-format production, not individual reading.
In short
- Convert documents in bulk and keep a shared history.
- A pronunciation dictionary the whole team shares.
- Embed audio in the LMS instead of distributing files.
- Documentation written for a conformance review.
What this does for you
The work is shared, so the tool should be
Most tools in this category assume one person reading their own material. Alternative-format production is a queue worked by several people. Shared history means the second coordinator can see that a document has already been done.
Fix a pronunciation once
Course codes, place names, and technical terms get mispronounced consistently. A workspace pronunciation dictionary means the fix applies to everyone's conversions, not just the person who found it.
Reduce the request queue
Embedding a listen button in the LMS converts a stream of individual requests into something students self-serve, which is usually a larger win than making each conversion faster.
Documentation you can hand to a reviewer
An accessibility statement that lists known limitations is more useful in a conformance review than one that claims none.
How to use it
Set up the workspace
Add your coordinators, import your pronunciation list, and set the default voice and language.
Work the queue
Convert documents, check extraction, and publish. History is shared, so nothing is converted twice.
Move upstream
Where a course generates repeat requests, embed the widget on the page so students do not need to ask.
An example
A two-person office handles 60 alternative-format requests a term. They add 30 course codes to the pronunciation dictionary in week one. By week four they have embedded listen buttons on the three modules that generated half the requests, and the queue for those modules stops.
Accessibility
- We publish known limitations, including no OCR for scanned documents.
- Keyboard operation and screen-reader support are treated as defects when they break.
- The free tool on this site is usable without an account, so you can test before asking anyone to sign up.
- We will answer specific conformance questions directly rather than referring you to a badge.
Questions people ask
Does this replace a human reader?
No, and it should not be sold as one. Synthetic speech handles prose well and handles mathematics, complex tables, and figure descriptions poorly. It removes the routine conversions so human effort goes to the material that genuinely needs it.
How does this relate to WCAG?
Audio alternatives are one route to satisfying requirements around text alternatives and readable content, but generating an MP3 is not itself conformance. What it does is make the alternative-format part of your obligations tractable at volume.
Can we bulk-convert a whole module?
Documents are converted one at a time in the app. For genuine bulk work, the API on higher plans is what teams use to process a term's readings in a batch.
What about mathematics and scientific notation?
Poorly, and we would rather say so. Inline notation is read literally and MathML is not interpreted. Material that is mathematics-heavy still needs a human.
Does it handle scanned documents?
No. There is no OCR. If a large share of your queue is scanned material, you need a tool with OCR alongside this one — NaturalReader is the usual answer.
Try it now
The converter on this site is free and needs no account. A 7-day trial adds document upload and longer text.
Start free trialNeed help? Contact support