Counsel.day began with one unanswered question.
In the autumn of 2024, James (41) and Alex (35) could not agree on whether to keep raising their son Tom in a small village in the countryside or move into the city for his first year of primary schooling. They had been talking about it on and off for nearly a year. The conversations went in circles. The first build of Counsel.day was the engineering attempt to break the circle: thirty evenings, one private vote each, sealed until the last day. This page is how the question became a product, and the rules we hold while we build it.
A question a single dinner could not resolve.
James and Alex had been talking about Tom's schooling for nearly a year. The village school was small, twelve children to a class, a teacher they both liked. The city school was larger, more academic, twenty minutes by car on a clear morning, fifty minutes on a Tuesday. James leaned toward the city for the academic shape of it; Alex leaned toward the village for the slow childhood. Neither of them was wrong. Every conversation about it ended in the same place: with both of them tired and the question still on the table.
The first build of the product ran for thirty evenings on whether to move. The verdict on day thirty was a Lean Yes, with a synthesis that named what neither of them had said out loud: the question was less about academics and more about the kind of family they wanted to be on a Tuesday. They read the verdict together on the Saturday morning and made the decision the same week. The first paid decision the product ever processed was someone else's, three months later.
The observation underneath the build was small and specific: a real decision a couple is carrying does not behave like a debate. The position you hold on Monday evening is not the same position you hold on Friday morning. The argument that won the last conversation is the argument the other person has rehearsed all week. The data of a long decision is the trajectory of two private feelings across time; almost none of the conversational tools available to a couple capture that data. Counsel.day captures it.
Rules that hold the work, even on a hard week.
Six beliefs hold while we build. They are pinned to the wall in plain English; they survive contact with the next feature decision. They are the test for whether we keep something.
Counsel.day does not tell you what to decide. It holds structure around the thinking, makes the data visible at the end, and names the axis underneath. The verdict is an input, not an instruction.
If the votes could be seen early, the verdict would not be honest and the product would not exist. The seal is enforced at the database, not the application; even with full administrative access, we cannot bypass it.
A decision that needs a season cannot be resolved in a single sitting. The duration is the second decision the product asks of you, after the question. Match it to the weight.
You pay for the verdict, not the days. The full five-layer analysis is the thing you take away; the daily votes are the means. The product is priced per decision because that is the unit of value.
We do not coach, advise, counsel, or therapise. We build a clean container, lock the seal, write the verdict honestly, and stay out of the deciding. The boundary is the product.
No jargon in the verdict; no clinical posturing in the marketing; no euphemism in the pricing. Six dollars is six dollars and a Lean Yes is a Lean Yes. If we cannot say it plainly, we have not understood it yet.
A data professional, not a clinician.
James Graham is a data professional. Twenty years building decision systems for organisations that needed to be honest with themselves at scale. No clinical training. No counselling qualification. No therapeutic credential. Counsel.day is what someone who has worked with data for twenty years built when they ran out of patience for a question their own household could not resolve. The full biography lives at James Graham · founder; the short version is on this page.
The team is small. The work that requires clinical thinking, we do not do. The work that requires data thinking · the seal, the trajectory, the agreement curve, the theme extraction, the synthesis prompt iteration, the privacy mechanism · is the work we do.
The name is the product, said in two parts.
Counsel is the older word for thoughtful weighing: the act of seeking and giving counsel, considering a question with care, arriving at a position rather than being told one. We did not want a name with advice, coach, or therapy inside it, because those words imply a position handed down from outside. Counsel implies a position arrived at from within. The product is built around that distinction. We hold structure around your thinking; we do not tell you what to decide.
.day is the mechanism, said as the domain. The product runs day by day, evening by evening, across the duration the question deserves: a week, a month, a season, a year. It is the opposite of a single-sitting tool. The name says, in three syllables, what the product is and how it works: counsel, every day, until the question has been given the time it deserves. The .day TLD was a deliberate choice over .com for that reason; the suffix is part of the meaning, not an accident of availability.
The honest list of things Counsel.day does not do.
One of the cleaner ways to describe a product is to name what it is not. The list below is what Counsel.day is not, in plain language, written so that anyone arriving from a search for a different category of tool can quickly route themselves to the one they actually want.
Counsel.day is not couples therapy, not individual therapy, not a therapeutic intervention. We have no clinical training and the product is not validated, endorsed, or tested by clinicians. If you need a therapist, the comparison page lays out the difference; find a therapist before you start a decision if the pattern is the problem rather than the question.
Counsel.day does not tell you what to decide. The verdict names where the data points; the deciding stays with you. If you wanted a recommender, there are plenty; we are not one.
No coach, no facilitator, no human in the loop on your decision. The pipeline runs in code; the synthesis is written by an LLM with no identifying metadata. No human at Counsel.day reads your votes, your notes, or your verdict.
Counsel.day is not Day One, Stoic, or Reflectly. It captures structured data across two or more participants, not free-form reflection from one. Many users keep both habits.
No streak. No badges. No push notification that wants you to come back. One evening prompt, optional, gentle, and the rest of the day is yours.
Counsel.day is not a crisis tool. If you or someone you love is in distress, the right thing is to reach a national helpline or an emergency service. We are a decision tool; the timescales we work on are days and weeks, not minutes.