AI disclaimer
The short version
- Wordaya is for a parent or guardian to use with their child. It is not designed for a child to use alone; please stay with them.
- Every explanation is generated by an AI model, on the spot, for that one word.
- It can be mistaken. It can be out of date. It can miss the meaning your child actually meant.
- It is a starting point for a conversation with your child, not an authority.
- It is not advice from a doctor, a lawyer, a teacher or any other professional.
- If an answer looks wrong or upsetting, tell us: we read every report.
Using Wordaya with your child
Wordaya is meant to be used by an adult, together with a child. The account must be held by someone aged 18 or over, and it is built on the assumption that the adult is nearby.
You do not have to hand over the keyboard. A perfectly good way to use Wordaya is for you to type the word, read the answer yourself, and then decide what to do with it: say it again in your own words, or press Listen so that your child hears it read aloud. Nothing is spoken until that button is pressed — the answer always appears as text first, and it stays on screen while the audio plays, so you can read it before your child does.
- Sit with them, at least while they are getting used to it. The app reads answers aloud, which is what makes it usable by a child who cannot read yet — and also means an answer can arrive without anyone having read it first.
- Use the Parent Zone. Every word your child looked up and every answer they received is listed there, newest first, and can be downloaded. That view exists for exactly this reason: so that you can see what your child has been reading.
- You decide what is appropriate. Wordaya is not a filter and not a babysitter. It has no idea what your child is ready for; you do.
- Some words open difficult conversations. Illness, death, bodies, bullying, divorce. When a word leads somewhere hard, that is a conversation for a grown-up rather than for a screen.
- If something looks wrong, tell us. Reports are read by a person and they change the instructions we send to the model.
What the AI is, and where it comes from
The explanations are produced by third-party AI models, which we call as a service. We do not train those models, we cannot see inside them, and we cannot make them say only true things. What we do is build the app around them: we choose the instructions they are given, the length, the language, the age level, and the safety settings.
That means a few things you should know:
- The same word can produce different answers. There is no fixed dictionary entry behind the app. Pressing "Try a different explanation" usually gives you a genuinely different one, which is useful, and also a sign that answers are not guaranteed.
- Answers get worse as questions get vaguer. A word with several meanings — "bat", "bank", "light" — may be explained in the sense your child did not mean.
- Every language is a translation away from a mistake. Wordaya supports twelve languages. An answer in a language other than English may read less naturally, and may be less reliable.
- Recency is not guaranteed. The explanation is written from what the model learned, not from a live check of today's news, law, prices or science.
What we do to make answers better
- Each explanation is pitched at the age step you chose, and written to be short enough that a child will actually read it.
- The instruction given to the model asks for plain language, no padding, and no talking down to the reader.
- The provider's safety filters are left on, so the model declines the material it is built to refuse.
- We keep no incentive to hold your child's attention. There are no adverts, no streaks, no notifications, and nothing that rewards coming back.
- Reports from parents are read by a person, and they change the instructions we send to the model.
What it is not
- It is not professional advice, and must not be used as if it were: not medical, legal, financial, psychological, or safety advice.
- It is not a citable source. Nothing here should be quoted in school work, a legal document, or a translation you need to rely on, without checking elsewhere.
- It is not a substitute for a person. For anything that matters — a health worry, a safeguarding concern, a decision about your child — talk to a qualified human.
- It is not supervised while it runs. No human reads the answer before your child does.
Sensitive words, and what to do about them
Children look up all sorts of words, including uncomfortable ones: illness, death, war, bodies, bullying. The models have safety filters, and we ask for careful, age-appropriate language, but we cannot promise that every answer on a difficult subject will land well.
If an answer upsets your child, or contains something that should not be shown to a child, please tell us at wordkeeper@wordaya.com with the word, the language and roughly when it happened. If it is urgent, or a safeguarding concern, please contact the appropriate service in your country as well — we are not able to act on anything urgent.
How to check an answer
- Read the history. The Parent Zone shows every word your child looked up and the answer they received, newest first, and you can download all of it as a CSV file.
- Ask your child what they understood. The gap between a correct answer and a misunderstood one is usually where the interesting conversation is.
- For anything that matters, check a second source: a dictionary, a textbook, a teacher, or someone qualified in the subject.
The legal bit, in plain words
We provide Wordaya as it is, without a warranty that an explanation is accurate, complete or suitable for any particular purpose. We are not liable for decisions taken on the basis of an AI-generated explanation. Nothing on this page limits the rights you have as a consumer under the mandatory law of your country, or our liability where the law does not allow it to be limited.
Contact
Reports, questions and corrections: wordkeeper@wordaya.com
Wordaya is an independent project. The AI models are provided by third parties; the explanations are generated for you at your request, and we do not claim authorship of them.