How it works

Clarity asks a few short questions and builds your summary with fixed, reviewed rules that run on your own device. There is no AI, and no one reads your answers.

What it can and can't do

• It can help you separate what you know from what you fear, and choose one next step.
• It can't tell you what anyone else thinks, feels or will do. It gives no scores and no verdicts.
• It's a reflection tool, not therapy, and it can't respond in an emergency.

Why am I seeing this?

Every theme in your summary lists the answers behind it. A theme only shows when several of your answers point the same way.

Safety comes first

Everyone is asked a few gentle safety questions, and you can skip them. If something you share suggests you may not be safe, support services come first, and conversation tips and offers are switched off.

How is this different from asking a chatbot?

• Safety questions come first, every time.
• No verdicts on your partner, and no guesses about their feelings.
• A structured Facts vs Fears worksheet instead of open-ended chat.
• Your answers aren't sent anywhere, and you can see why every result appears.

Where the ideas come from

The tools draw on everyday approaches to worry, reassurance and disagreements. Clarity itself hasn't been tested in research. Where the research is small or early, we say so.

• Writing down a thought and the evidence for and against it: the NHS guide Reframing unhelpful thoughts; McManus and colleagues, 2012.
• Asking for reassurance again and again: Kobori and Salkovskis, 2013; Starr and Davila, 2008.
• What helps or hurts during disagreements, such as hostility and pulling away: Woodin, 2011; Lavner and Bradbury, 2012.
• Why your answers aren't turned into a label or category: Fraley and colleagues, 2015.
• Saying how you feel, rather than what they did: Rogers and colleagues, 2018 (weak evidence).
• Slow breathing with a longer breath out: Fincham and colleagues, 2023; Balban and colleagues, 2023 (early findings).
• Setting aside a short time for worry: Borkovec and colleagues, 1983; Versluis and colleagues, 2016 (mixed results).
• Writing about your feelings: Frattaroli, 2006 (small effects).
• Checking a partner's social media: Muise and colleagues, 2009.