Limited early access

ArticScore access and pricing

ArticScore is in limited early access, rolled out case by case, and we do not publish a price list yet. Access is arranged directly so that we can size the commercial terms against your volume, your sounds and age ranges, and your child-privacy requirements before anything is turned on.

What we will ask you

  • What you are building, and which sounds and age ranges matter most to you. The engine performs unevenly across sounds, so this is what determines whether it fits your product.
  • Expected volume. “Not sure yet” is a perfectly good answer at this stage — it changes what we quote, not whether we reply.
  • Your child-privacy posture. Whether you are consumer or school-deployed, what consent mechanism you use, and what retention you need. This shapes the written terms more than the price does.
  • What you are using today, if anything, and what is not working about it. It tells us more about the requirement than a feature list does.

What early access includes

  • The scoring endpoint, returning per-phoneme scores, the sound heard instead, error type, confidence tier, and timing spans.
  • The current specification, and notice ahead of schema changes rather than after them.
  • Written data-processing terms covering retention, training use, and subprocessors.
  • A direct line to the people who built it. Early integrators shape what gets fixed first, and at this stage that is a real part of what you get.

Request access

We reply to every request. Include your use case and your expected volume and the first reply can be a useful one.

A rough number changes what we can quote, not whether we reply.

ArticScore is in limited early access, rolled out case by case. We reply to every request.

Looking for the therapy platform rather than the API?

SpeechTherapyMagic itself — games, activities, and pronunciation practice for children, used by SLPs and families — has public pricing and a free plan. ArticScore is the developer API underneath it.

Access questions

How much does ArticScore cost?
We do not publish a price list. ArticScore is in limited early access and pricing is arranged directly, because at this stage the right commercial shape depends heavily on volume, retention requirements, and the sounds and ages you need covered. Request access and we will come back with specifics.
Why is there no public pricing?
Because the product is in early access and the schema is not yet frozen. Committing to a public price would mean committing to a specification we are still improving, and to a set of terms that vary considerably with volume and with the child-privacy requirements of the deployment. Access goes out case by case for the same reason.
Is there a free trial?
Request access and tell us what you want to test. Early access is arranged case by case rather than through a self-service signup, and evaluating the engine on your own audio before committing to anything is exactly what we would want you to do.
Can I evaluate it against my current provider?
Yes, and we would encourage it. Our published benchmark reports our own measurements against human judges on a public corpus, including the results that go against us — but a benchmark on someone else's audio is not a benchmark on yours. Our buyer's guide sets out a protocol for running the comparison on your own recordings, and it is the one we would follow ourselves.
What languages does ArticScore support?
American English is the evaluated and supported dialect today. If you serve another variety of English, that is worth raising early — patterns like th-fronting and dark-/l/ vocalisation are standard in some varieties and would be flagged as errors against an American English lexicon.

Before you get in touch

Accuracy benchmark

The numbers, the caveats, and where we lose.

What the API returns

The full response schema, field by field.

Children's privacy and COPPA

What the rule requires before you record a child.

Choosing a pronunciation scoring API

What to test for, and how to run the comparison on your own audio.