← GreyRocked

How GreyRocked comes up with its answers

Where the replies come from, what shapes each one, and the research behind the ideas, including what has not been tested. If someone may see your screen, close this page now.

First: are you safe?

In danger right now? Call 911. UK: 999. Europe: 112. Australia: 000. Elsewhere: your local emergency number. If you can't speak, do not hang up.

UK: call 999, and if you cannot speak, press 55 when you hear the recorded message. In many US areas you can text 911.

If someone in your family hurts you, scares you, or controls you: your safety comes before any reply from this app. The time right after you tell them you want less contact can be the most dangerous.

If they have been violent, don't tell them first: make a plan with a hotline (a free phone line). Don't reply to a threat to hurt you. A threat to hurt or kill you can be reported to the police. Save the messages (a screenshot is fine).

If you are the one thinking about ending your life: if you have already hurt yourself or taken pills, call 911 (UK: 999, Australia: 000) now. If you are thinking about it, call or text 988 (US and Canada). UK: Samaritans, 116 123. Australia: Lifeline, 13 11 14. Elsewhere: findahelpline.com.

Be honest first

GreyRocked is a playbook built on ideas from published psychology research. It is not a clinical treatment, and the app itself has not been tested in a trial. The "grey rock" method has no controlled studies of its own that we could find. What follows is the research behind the pieces the app uses, and what each study did and did not look at. No clinician has reviewed the app's wording yet.

What it does

Nobody decides for you, and nobody is made the bad guy. You are always the person being hurt. The app never blames you, and no setting changes that.

It is a written playbook, not a guess and not a chatbot. Every answer comes from a fixed set of scenarios (25 of them: a guilt trip, a deadline, a money ask, a fake apology, "it was only a joke", and so on), written ahead of time. Nothing you tap is sent anywhere, so the same situation gets the same careful answer every time, with no AI making things up.

The safety rule on every draft: each suggested reply is safe to be read aloud to the whole family. No reply is a trap, a threat, or a weapon.

Three things shape each answer

They are kept separate on purpose.

  1. The pattern. The app names what the message is doing (for example "guilt trip" or "deadline pressure"), because a pressure you can name is easier to say no to. It judges the pattern, not the person.
  2. Your posture. You pick how you want to handle this person (for example short and firm, or no contact). Your posture decides the action: what to say, or whether to say anything.
  3. Your lantern. The lantern is the set of values you hold (for example faith, peace, repair, protecting yourself first). It does not change what you do; it changes the reason, so the reply fits who you are.

Where the values come from

Every line that reasons from a tradition carries its source, so someone who holds that view can disagree with something specific, not in general.

The Jewish and Muslim lanterns (teshuvah, sabr) were written by people outside those traditions and still want a reader from inside them.

The psychology behind the ideas

1. Naming the pressure makes it easier to stand against.

The app names what a message is doing. In a brain-imaging study of 30 people, putting a name to an upsetting picture calmed the brain's alarm centre (the amygdala) compared with looking without naming (Lieberman et al., 2007, Psychological Science). In seven preregistered experiments (six randomized, 6,464 people, plus a field study on YouTube with 22,632), short videos teaching the named tricks of manipulation, like scapegoating or false either-or choices, made people better at spotting them (Roozenbeek et al., 2022, Science Advances). The effects were small to medium in the lab and small in the YouTube test. Those studies were about pictures and online misinformation, not family messages. We borrow the idea; we do not claim the app was tested.

2. Settling yourself first, then answering short.

A study of 30 married couples recorded their bodies' stress arousal while they talked about conflict. Three years later, among the 19 couples the researchers could find again, the more aroused a couple had been, the more their marital satisfaction had dropped (Levenson & Gottman, 1985, Journal of Personality and Social Psychology). That study measured arousal and satisfaction; it did not test what to do about it. Taking a pause to settle when you feel overwhelmed is a common recommendation from the same research group, and it is the idea behind the app's short, calm replies and its "not now". This is research on couples, not on families. It is a small study, and we are not claiming a number of heartbeats or any rule of thumb from it.

3. Deciding ahead of time what you will say.

The app gives you ready words and a posture chosen in advance. A review of 94 tests found that deciding in advance "if this happens, then I will do that" helped people carry out their goals more than good intentions alone, a medium-to-large difference (Gollwitzer & Sheeran, 2006, Advances in Experimental Social Psychology; d = 0.65). That review was about reaching personal goals in general, not about family conflict.

What this page does not claim. That grey rock "works" for everyone; that the app is clinically validated; that any therapist endorses it; or anything from psychology about values and wellbeing (only the traditions above are cited for the lanterns).

Try it, free

Tap who sent it and what they did, or paste their message in. What you type stays on your phone. No account, no signup.

Open GreyRocked, free 🌱

More help: Grey rock method examples · Setting boundaries with toxic parents