Decide by circle, not by person. Start with the number of invitations you can send, then work outward in a fixed order — immediate family, grandparents, the aunts-and-uncles and cousins layers, your parents' guests, the wedding party, close friends, friends' partners, the people who invited you, coworkers, neighbors, children — and take each circle whole: everybody in it, or nobody. Almost every guest list that hurts somebody was decided one name at a time, because a name is comparable and a circle is not. The decision tree below routes each household into a circle and a tier as you answer it, the builder records the answers as a list you can count, and each branch hands off to the page that settles it properly. Eighteen worked lists follow, and none of it leaves your browser.
This is the tree the wedding blogs publish as a picture you can only look at. Answer it here instead: work down in order, stop at the first question that fits the household in front of you, and the answer is not a verdict — it is a circle and a tier, which is a row on the sheet in the builder below. Run it once per household, not once per person, because an invitation goes to a household.
Circle: immediate family, grandparents or wedding party. Tier A. This is the only group nobody negotiates, and it is usually 15 to 30 people once both sides are counted. Do the whole of Tier A in one sitting, before anything else, because it is the number every later decision is measured against. If Tier A alone is over the number you can seat, the problem is the room and not the list.
Take the whole layer or none of it. Tier B, or not inviting. Family branches are layers, and a layer is the one thing everybody in it can see. Three of seven first cousins invited is four people comparing notes by the end of the day the invitations land. Set the boundary at a generation — "first cousins yes, second cousins no" — and run the identical boundary on both sides, however lopsided the totals come out. Two exceptions, handled below: a relative who is genuinely somebody else's guest, and a relative you are estranged from.
Circle: parents' friends. Counts against their allocation, not yours. Agree a number with each set of parents before a single name is written down — a third each is the common split — and let them spend it. That one conversation converts "can we add the Hendersons?" from a request you have to refuse into arithmetic they can do themselves, and it stops their additions from silently eating your own seats.
Circle: close friends, friends' partners or people who invited you. Tier B or C. Friends are individual calls; friend GROUPS are layers, because a group chat makes a partial invitation public within the hour. Pick one standard for partners — married, engaged or living together is the usual line — and apply it to every friend on the list rather than ranking your friends' relationships one at a time. An invitation you accepted three years ago is a reason to think, not a debt.
Circle: coworkers or boss. Whole team, or nobody from that team. Run the count before you run the feelings: if you would be inviting most of a team and leaving two people out, invite none of them or invite all of them. Your manager is a separate, named decision and never an automatic one. Work invitations are also the easiest circle to cut whole, because "we kept it to family and friends outside work" is a complete sentence nobody at the office can argue with.
One policy for children; a named decision for everyone else. Children are a whole-list policy — either they are on the list or they are not — and the first exception rewrites the policy for everybody. Estrangement is the opposite: it is a single named person, decided once, recorded once, and never disguised as a layer rule, because dressing it up as a boundary quietly cuts three relatives you did want there. Anyone you have decided against goes in the not-inviting pile, which stays on the sheet and stays out of the downloaded file.
Each branch has a page that finishes it, and the circles come in the order below. Immediate family and grandparents are Tier A and need nothing from this cluster. The family layers — aunts, uncles and cousins — are settled on do you have to invite extended family. Your parents' guests, and how many of them there are, is how many guests parents get to invite. A relative you are estranged from is the estranged-family decision. The wedding party and close friends, plus the friends you have drifted from and the couples whose weddings you attended, are friends and who invited you, and whether a friend's boyfriend or girlfriend is on the list at all is your friends' partners. Work is inviting coworkers, children are deciding on a child-free wedding, and if the whole thing still does not fit, cutting the list down takes layers off rather than individuals — or, if you are working to a genuinely small room, who to invite to a small wedding shows where the line falls at 20, 50 and 75. What you say to the people who did not make it is telling someone they are not invited. And the sheet all of this lands on is the free wedding guest list template.
Each card is one household taken through the six questions, ending where the tree always ends: a relationship circle and a tier, which is exactly what the builder stores. Copy one as plain text, or hit Personalize to open the builder on the page that settles that branch.
HOUSEHOLD: Elena & Marco Alvarez — your parents Q1 · Would the day be wrong without them? Yes. → Circle: immediate family · Tier A · 2 seats · 1 invitation No further questions. Tier A is the fixed cost of the wedding, and it is the number every later cut is measured against.
HOUSEHOLD: Iris Whitfield — your grandmother, plus a carer Q1 · Would the day be wrong without her? Yes. → Circle: grandparents · Tier A · 2 seats · 1 invitation · +1 (carer, unnamed) The plus one here is not a guest, it is access. Count it as a seat and stop thinking about it.
HOUSEHOLD: Camila Reyes — a first cousin you last saw at a funeral Q1 · Core? No. Q2 · Relative outside the core? Yes. → Circle: first cousins · the LAYER decides, not Camila Layer as it stands: Diego, Paloma, Nico in · Camila, Héctor, Lucia, Mateo out ⚠ 3 of 7 first cousins on the list and 4 not. → Either all seven go to Tier B (7 invitations, 9 seats) or all seven go to not inviting. There is no third answer that survives the invitations landing.
HOUSEHOLD: Peggy & Ed Nolan — your mother's oldest friends Q1 · Core? No. Q2 · Relatives? No. Q3 · Your parents' people rather than yours? Yes. → Circle: parents' friends · Tier B · 2 seats, counted against your mother's 30 Her allocation after this: 28 left to spend. Nothing came off your own number, which is the entire point of agreeing an allocation before anyone names a name.
HOUSEHOLD: Jonah Pike & guest — a college friend, partner of eighteen months Q1-3 · No. Q4 · Friend, and a friend's partner? Yes. Standard agreed once, for everybody: married, engaged, or living together. Jonah and his partner live together → both named · Circle: college friends · Tier B · 2 seats · 1 invitation The same test then runs on all six college friends without you having to think about any of them again.
HOUSEHOLD: Dana Ruiz — on your team of nine at work Q1-4 · No. Q5 · Work? Yes. Count first: inviting 6 of 9 leaves 3 people sitting next to 6 who were invited. ⚠ Partial team. → Whole team to Tier C (9 seats) or the whole circle to not inviting (0 seats). Chosen: the whole circle out. The sentence that goes with it: "we kept it to family and friends outside work." Nobody at the office can argue with that, and nobody has to.
HOUSEHOLD: Gerald Whitfield — an uncle nobody in the family speaks to Q1-5 · No. Q6 · Estranged? Yes. → Circle: estranged family · Not inviting · reason recorded once, in his own row The aunts-and-uncles layer above him is untouched and still consistent. The consistency check stays quiet, because a named-person decision is not a split layer — and that distinction is the difference between one hard conversation and four.
Once while you decide it, once when the invitations go out, and once when you chase the households that never replied. Build it once here, download the file, and every name after that is a copy rather than a retype — Text My Wedding reads this exact file and texts each household on it.
The eighteen circles the builder ships with, grouped the way couples actually work through them: the core, then the family layers, then the people somebody else is bringing, then friends, then work, then the two whole-list policies. Each card is the circle written out as rows, with the default rule and what it costs. Work down them in this order and you will never be asked to compare two people from different circles, which is where lists go wrong.
CIRCLES 1-2 · THE CORE — decided first, never trimmed Immediate family — parents, step-parents, siblings and their partners Grandparents — named individually on both sides, carers counted as seats Worked list: Elena & Marco Alvarez · Sofia Alvarez (+1) · Ruth & Alan Whitfield · Iris Whitfield (+1) — 6 people, 4 invitations, 8 seats, all Tier A. These are individual households and not a layer: nobody compares them, and nothing here comes off to make room.
CIRCLES 3-7 · THE FAMILY LAYERS — all or nothing, both sides identical Aunts and uncles — a layer First cousins — a layer, and usually the biggest block on the list Second cousins — a layer, and almost always cut whole Partner's relatives — mirror your own boundary circle for circle Estranged family — never a layer; one named decision at a time Boundary written down once: "first cousins in, second cousins out, on both sides." — Alvarez side: 2 aunts and uncles, 7 first cousins → 11 seats. Whitfield side: 1 aunt with a guest → 2 seats. The rule matched. The totals never will, and they do not have to.
CIRCLE 8 · PARENTS' FRIENDS — somebody else's guests, somebody else's number Agreed before any names: 120 seats, a third each. You two — 60 · Alvarez parents — 30 · Whitfield parents — 30 Spent by the Alvarez side: Peggy & Ed Nolan (2) · Rosa & Julio Alvarez (2) · Barbara Kent (not inviting) — 4 of their 30 gone, 26 to spend, and not one seat off your own sixty. The allocation is the whole trick: it turns every later request into their arithmetic rather than your refusal.
CIRCLES 9-14 · FRIENDS — individual calls, group layers, one partner standard Wedding party — everyone standing up with you, partner named, no argument Close friends — individual calls Childhood friends — a group, so a layer College friends — a group, so a layer Friends' partners — one standard applied to every friend on the list People who invited you — one decision per couple; an invitation you accepted is not a contract — 24 people, 14 invitations, all Tier B. The two "group" circles are the trap: half a group chat being invited is public information before you have finished addressing the envelopes.
CIRCLES 15-17 · WORK AND NEIGHBORS — count before you feel Coworkers — a team is a layer. Most of a team invited and two people not is the most visible cut on any list. Boss or manager — a named person, decided separately, and never because of the reporting line alone. Neighbors — the people who will see the cars arrive. Usually all of the immediate ones, or none. Team of 9 · would invite 6 · would exclude 3 → ⚠ partial team — Whole circle to not inviting: 0 seats, one sentence, nothing to explain on Monday.
CIRCLE 18 · CHILDREN — the one circle that is a whole-list policy Policy options, chosen once, applied to every household: — No children at all — Immediate family's children only — Over-12s only — Ceremony only Chosen: no children at all. 22 children across 12 households → 0 seats, 22 recovered. The first exception is not an exception; it is the new policy, and every household you told "no" will find out about it.
Nobody publishes this half, which is why it is the half people get wrong. These are the five groups that come off almost every list, written as the rule rather than as a judgment about anybody — plus what the not-inviting pile is actually for. Each one stays on the sheet, counted nowhere, and out of the file you download.
NOT INVITING · PEOPLE YOU HAVE NOT SPOKEN TO IN A YEAR The test, applied to everyone equally: have we exchanged more than a birthday message since last summer? No → not inviting, whoever they are. — 11 households · 17 people · 17 seats recovered. The test does the cutting, so no individual name has to be defended. Run it on your own side and your partner's on the same evening.
NOT INVITING · THE WHOLE OF A LAYER YOU CANNOT AFFORD WHOLE Second cousins — whole circle out First cousins once removed — whole circle out Great-aunts and great-uncles — whole circle out — 3 circles, 1 boundary, 0 individual decisions. "We stopped at first cousins" is an answer. "We ran out of room" is an opening to negotiate, and somebody will take it.
NOT INVITING · A TEAM YOU CANNOT INVITE ALL OF Team of 9 · you would genuinely want 6 → Invite 6 and 3 people spend a Monday knowing. Invite 9 and you have added 9 seats to a list that did not fit. Chosen: none of them. — 9 seats recovered, one sentence, no ranking of colleagues written down anywhere.
NOT INVITING · SOMEBODY WHOSE PRESENCE WOULD CHANGE THE DAY FOR SOMEONE ELSE One row · one named person · one reason, written down once Not a layer rule, not a capacity decision, and not something to re-argue every two weeks. — The circle above them stays intact and stays consistent. The consistency check does not fire, because this was never a boundary question.
NOT INVITING · THE PEOPLE YOU WERE ONLY INVITING BACK Invited you to their wedding in 2023 · you have seen them twice since → Reciprocity is a reason to think about somebody, not a debt that transfers. Chosen: not inviting · 2 seats recovered. If it helps: they will almost certainly have made the same call about somebody, and nobody sends an invoice.
This is the distinction the whole cluster turns on, and the one the incumbent guides skip. A call about one named individual, a boundary around a whole circle, a policy over the entire list and a split between the people paying are four different objects. Almost every guest list argument is two people applying rules at different levels and both being right — name the level first and most of it evaporates.
The whole list · The list as a whole. Enforced name by name it becomes six weeks of trading one relative for another; enforced by circle it is three decisions that hold. This is the number the builder asks for first, and every rule below is applied underneath it.
A whole layer · Every household in the circle. Applied to one person it becomes a favor, and the rest of the layer finds out what the favor was within a day of the invitations arriving. This is the rule the consistency check is watching.
A whole layer · Every household in the circle. Half a team is the most visible cut on any list, because the people left out sit next to the people who were not. Cut the team whole and the sentence writes itself.
The whole list · The list as a whole. One exception re-opens the policy for everyone, because the exception is the new rule. Either it is a whole-list policy or it is a series of favors you will be defending until the day itself.
One person · One named individual. Dressed up as a layer rule it quietly cuts three relatives you did want there, and everybody in the family knows which decision it really was. Give it its own row in the not-inviting pile and leave the layer above it intact.
An allocation · Each contributor's bucket. Left unstated, every name your parents add looks like a cut somewhere else on your own list. Stated up front, it is arithmetic they can do themselves.
This page decides who is on the list. Everything downstream of that decision belongs somewhere else and is linked rather than repeated. The number the builder asks for first comes from the wedding attendance rate — it turns the room you have booked into the number of invitations you should send, and that answer is exactly what the capacity field wants. Once you have the names, collecting everyone's contact details is the next job, and this builder never asks for a phone number or an address. How many seats a partner is allocated is settled by the plus-one rules, how to word the child-free decision on the invitation itself is adults-only invitation wording, and a layer you cut that might come back later is timed on second-wave invitations. What one more name costs at the bar is worked out by the alcohol calculator, and if you are building around a traveling list, the destination wedding timeline sets the dates around it. One boundary worth naming: the wedding planning spreadsheet is a checklist of tasks and a budget, not a guest list. Keep them in two files.
Etiquette on a guest list is not about who is owed an invitation — almost nobody is. It is about whether the same rule was run on everybody, and whether the people it excluded heard it from you rather than from somebody else.
Not your cousins, not your colleagues, not the couple whose wedding you went to last year. The obligation people feel is real and the obligation itself is not. What you do owe is consistency, and telling the people it affects yourself.
Open the circle first, agree whether it is in or out, and only then read the names. Reversed — names first, circle after — you are comparing individuals, and every comparison is one you will be asked to justify.
Whatever circle you stop at, your partner stops at the same one. The totals will be lopsided and that is fine: people compare the rule, not the number of seats each family got.
Set the standard once — married, engaged or living together is the common one — and apply it to every friend on the list rather than deciding relationship by relationship. Whoever is on the invitation is named on the invitation.
Anyone who will definitely notice — a sibling's best friend, a cousin in a layer you cut, a colleague who has been asking — hears it from you, once, privately, in the two weeks before anything is mailed. Never in a group chat, and never as a maybe.
A decision that lives only in a conversation gets re-litigated every time somebody brings the name up. Keep the not-inviting pile on the sheet with its circle attached, so you can see at a glance that each layer came off whole.
None of these are catastrophes and all of them are avoidable in the hour before you start. Each one costs a conversation you did not need to have.
A list with no capacity behind it comes out bigger than the room, whatever the room holds, because nothing in it ever stops you adding one more. Put the number in first even if it is a guess — a guess you can revise beats an open list you have to cut later, with an audience.
Every individual decision invites a comparison, and every comparison is one somebody can take personally. Circles cannot be taken personally, which is the only reason they work.
The most natural instinct on the whole list and the most reliably damaging. The other four find out within a day, and the reason they would be given — closeness — is the one reason that cannot be said out loud.
"Just a couple more" arrives four separate times, and every one of them comes out of your own seats. Agree the allocation before anyone writes a name, and their additions become their arithmetic.
The exception is the new policy the moment somebody else hears about it, and somebody else always hears about it. Pick a policy that includes the household you were going to make the exception for, or hold the line.
A group message about who is not invited turns a private decision into a public event and hands everybody in it an audience. Tell the people it affects individually, before the invitations arrive.
A list that only exists as a conversation cannot be counted against the room, cannot be checked for a split layer, and has to be rebuilt from memory on the evening you finally sit down to send something.
The sheet starts empty and every circle is armed. Put the number of invitations you can send at the top, then add a household, name everyone on that invitation, and give the row the circle and the tier the tree above just handed you. The count and the envelope total move as you type, the circle check tells you the moment a layer has gone out half-invited, and the download is a plain spreadsheet. Everything runs in your browser — nothing you type is stored or sent anywhere, and there is no saved draft, which is exactly why the download matters.
Every circle armed and nothing tagged yet. Answer the questions above, and each name lands in a circle and a tier as you go.
This is how many people you can invite, not how many will come — turning one into the other is what the attendance-rate page does: to seat 100, invite 118 to 128.
Pick the circle a new name belongs to — it is what the consistency check reads.
0 people in 0 householdsEveryone in tiers A, B and C is counted here; 0 people are in the not-inviting pile and stays out of the file. Set a capacity to see where the line falls.
Nothing on the list yet. Add your first person, then give them a household — everyone on one invitation shares the same household label.
Everything here runs in your browser. Nothing you type is stored or sent anywhere — there is no account, no saved list and no cookie holding your guests (we count anonymous, bucketed usage of the tool itself, never a name, a number or a list size). That also means closing this tab loses the list, which is why the button says download.
Where the line falls
Downloads a CSV of everyone still on the list — tiers A, B and C, minus any tier you have switched off above. The not-inviting pile stays on this page.
Work outward in circles rather than through a list of names: immediate family and grandparents first, then the wedding party, then the family layers — aunts and uncles, first cousins — then your parents' guests out of an agreed allocation, then close friends and friend groups, then partners of friends under one standard, then work, then neighbors. Take each circle whole, because a circle cannot be taken personally and an individual decision always can. Stop when you reach the number of invitations you can send.
Start with a number, not a name. Work out how many invitations you can send, then run every household through the same six questions: is the day wrong without them, are they a relative outside the core, are they somebody your parents want there, are they a friend or a friend's partner, do you know them from work, and is this a child or a person you have already decided against. Each answer is a relationship circle and an A, B or C tier rather than a verdict — and if the total does not fit, you take whole circles or a whole tier off, never individual names.
The groups that come off almost every list are people you have not spoken to in over a year, a family layer you cannot afford to invite whole, a work team you cannot invite all of, anyone whose presence would change the day for somebody you love, and people you were only inviting back out of obligation. Cut each of them as a rule applied to everybody rather than as a judgment about one person, keep them written down as not inviting so the decision is not re-argued, and tell anyone who will obviously notice before the invitations arrive.
There is one on this page and it is interactive rather than a picture. Every other version of this chart hands you a verdict — invite them, do not invite them — and then you are on your own. This one routes each household into a relationship circle and a tier instead, which is a row you can count, check for split layers, and download. Answer the six questions above, tag the households in the builder as you go, and the chart's output is a spreadsheet rather than an opinion.
Almost nobody outside your immediate family is owed an invitation, so etiquette here is about consistency rather than obligation. Run the same boundary on both sides of the family, invite whole circles rather than picking through them, set one standard for partners and apply it to every friend, keep children as a whole-list policy rather than a per-family favor, and tell anyone who will notice themselves, once and privately, before anything is mailed. Nobody minds a small wedding. People mind being the exception.
As many as your room and your budget allow, and the honest order is room first, list second. Whatever that number is, it is a number of invitations rather than a number of names — some households are two people and one envelope, and some of the list will say no. Put the figure in the builder before you write a single name; a list built against a number is a series of small decisions, and a list built without one is one very long argument in March.
Yes, but over a number rather than over the list. Give anybody who is contributing a fixed allocation of their own and their guests come out of a figure they agreed to rather than out of yours, which turns every later request into arithmetic they can do themselves. The splits that hold up, and what to say when one family uses far less of its allocation than the other, are worked through on the parents page below.
Build the list here, download it, and import the file into Text My Wedding — the same names then carry the invitation, the RSVP and every reminder, without anybody retyping them.
