Random team generator
Paste your names and split them into balanced teams — either a fixed number of teams, or teams of a fixed size. Saved in your browser for next time.
0 entries. Blank lines are ignored, and repeated names are kept — two people really can share one.
Teams
The teams will appear here.
Save this list
Nothing saved yet. Lists you save stay in this browser — they are never uploaded.
How it works
Two things have to be random here, and most tools only get the first one right.
The first is who ends up with whom. The list is shuffled with the Fisher–Yates algorithm before anyone is placed, so every possible arrangement of people into teams is equally likely. That part is well understood and hard to get wrong once you have a correct shuffle.
The second is subtler and it is the one worth explaining. When the number of people does not divide evenly — nine people into two teams, say — one team has to be bigger. The obvious way to write that is to hand the spare person to the first team, then the second, and so on. It produces a perfectly valid split every time, so nothing ever looks wrong. But run it every week and Team 1 is the biggest team, forever. That is a real and permanent bias in a tool whose entire purpose is being fair, and it is invisible in any single use.
So this tool randomises which team gets the extra member, independently of who is on it. With nine people in two teams you get a five and a four — and which of them is the five changes each time you press the button. There is a test that runs thirty thousand splits and checks that each team position ends up largest about equally often, alongside a deliberately broken version that the same test has to reject. A check that has never been shown failing is not worth much.
What "teams of four" actually means here
If you ask for teams of up to four from thirteen people, this tool gives you four teams of 4 / 3 / 3 / 3 rather than 4 / 4 / 4 / 1. Both are defensible readings of the request. We chose the balanced one because the alternative strands somebody on their own, and in a classroom or a workshop that is not a rounding detail — it is the entire outcome for that person. The number you type is treated as a maximum team size, and nobody is left alone unless the list genuinely has only one person in it.
Worked examples
A class splitting into project groups
Thirty-one students, groups of four. You get eight groups, seven of four and one of three, with the smaller group in a different place each time you run it. Save the register once as “Year 9 Science” and the following week is two clicks rather than five minutes of typing.
Five-a-side when eleven people turn up
Two teams from eleven means one side plays with six. Choosing which side that is by hand is exactly where arguments start, and doing it in a fixed order means the same shirt colour is a player up every single week.
Pairing people up
Set the team size to two and you have a random pairing tool — for code review partners, language practice, or a secret-buddy scheme. With an odd number of people one group of three appears, in a random position.
Breakout rooms
Paste the attendee list, choose the number of rooms, and copy the result straight into the chat. The copy button gives you plain text with a heading per team, which pastes cleanly into most meeting tools.
Questions people actually ask
Can I keep two people apart, or together?
Not yet — and that is worth saying plainly rather than leaving you to discover it. Constraints like “these two are always on opposite sides” change the problem quite a lot: with enough rules a valid arrangement can become rare or impossible, and a tool that quietly reshuffles until it finds one can spin forever. Doing it properly means detecting that case and explaining it, which is a real piece of work rather than an afternoon.
Are the teams balanced by skill?
No. This is a random splitter, not a rating system. Everyone is equally likely to be on any team, which is the correct behaviour when what you want is fairness rather than competitiveness.
What happens to duplicate names?
They are kept. Two people in a group really can share a first name, and quietly merging them would drop somebody from the teams without telling you.
Are my names sent anywhere?
No. The split happens in your browser, and saved lists are stored on your own device. Nothing is uploaded and no server here ever sees a name. Open your browser's network tab and press the button — nothing goes out.
Can I ask for more teams than I have people?
Yes, and you will get empty teams, with a note saying so before you press the button. It is usually a typo, so the tool shows you rather than silently giving you fewer teams than you asked for.
Limitations, stated plainly
- No constraints between people — no “keep these two apart”, no “always put these together”. See the question above for why that is a bigger feature than it looks.
- No skill or seeding weights. Every arrangement is equally likely, by design.
- Saved lists live in one browser. They do not follow you to another device, because there is no account and nothing is stored on a server — which is also why the privacy claim above is true.
- Clearing site data clears your lists, and private windows forget everything when they close. The splitter still works; the saving does not, and the page says so rather than failing quietly.
Related tools
- Spinning Wheel
Spin a wheel over your saved list, with names crossed off as they are drawn.
- Random Order
Shuffle a list into a running order you can copy out.
- Dice Roller
Roll any number of dice from four to a hundred sides, with a running history.
- Coin Flip
Flip one coin or a hundred, and keep the tally.
- Random Number
Draw numbers from a range, with or without repeats.