In My Words

Say it
your way.

A short check-in a young person fills out right before an appointment and sends to their therapist or psychiatrist.

Saying how you're really doing is hard, especially to a clinician you might barely know, and a lot gets lost between how you'd put it and what they ask. Here you write it the way you'd actually say it. AI pulls out a few key points, you check every line, and your clinician reads it before you walk in.

A young person's tangled thoughts, written on a phone, become a single clear thread carrying their own words to a clinician holding the other end.“““
“I wasn't completely honest when I first filled out like the tablet… if they asked me if my mood was like 1 through 10, say it was like a 1 I said it was like a 3 or 4.”
A young person in a US study of depression screening (Radovic et al., 2021)

Write it, check it, send it

Write

Answer a few of the questions, or just write freely. Short answers are fine, and any question can be skipped.

Check

You see the summary before anyone else does, with your exact words under every line. You can remove anything or add a note.

Send

Your clinician reads it before you meet, and after the visit your plan comes back to you in plain language.

A brief you can check line by line

Every phrase in the brief is tied to the young person's own words, and anything ambiguous is flagged for you to ask about. After the visit, you approve a plain-language version of your plan before they see it.

How clinicians use it

See every question and why it's there

Each screen appears exactly as a young person would see it, with notes explaining the psychology behind the wording and the safeguards built around the AI.

Open the questionnaire

What people ask

For young people, parents, clinicians, and anyone curious about how it works.

What is In My Words?

It's a short check-in a young person fills out right before an appointment with their therapist or psychiatrist. They write about how things have been in their own words, AI pulls a few key points from the longer answers, and the clinician reads it before the visit. The young person sees exactly what will be sent first.

Is an AI going to talk to me?

There's no chat at all. People wrote the questions, and everyone sees the same ones. The AI's only job is to organize what you wrote into a short summary, and it doesn't reply, give advice, or diagnose anything.

Who sees what I write?

Only the clinician you're seeing. Before anything is sent, you see exactly what they'll get and can edit or remove answers. The one exception is anything about safety, which stays in. Pictures you remove are deleted.

What if someone writes about wanting to hurt themselves, or being hurt?

That part stays in the summary, and young people are told so before they write anything. The clinician may need to act on it. The page also makes clear that nobody is watching in real time, and it points to 988 for anyone who needs to talk to someone right now.

Will parents or guardians see it?

The tool sends the summary to the clinician and no one else. Whether a parent gets involved after that depends on the situation and on state law, the same as with any visit with a young person, and the privacy section explains this in plain language.

Can the AI get it wrong?

Yes, which is why everything it writes can be checked. Each line shows the young person's exact words, and code confirms that every quote really appears in what they wrote. When a phrase could mean more than one thing, it gets flagged for the clinician to ask about instead of being guessed at.

Why use AI at all?

A lot of young people find it hard to say how they're doing out loud, especially to someone they've just met. Writing first can make that easier, and the AI helps shape the writing into something a clinician can read in a minute while keeping the young person's voice intact.

Does it replace talking to a therapist?

No. It's meant to make the conversation a little easier to start. Nothing it produces is a diagnosis or a risk score.

Is this ready to use?

Not yet. It's a pilot built for a class research project, and it hasn't been tested or approved for real clinical care, so please don't enter real personal information.

How was it built and tested?

It follows a five-part, risk-based framework for AI in mental health care from an accompanying policy brief. It was tested on 20 fictional young people whose words are adapted from real, cited quotes in published research. No real teen data was used.

See it from both sides

Fill it out as the young person, then switch over to the clinician's screen and see what arrives.

Try the demo