Seeded random team generator

Enter the people, choose how many teams you need, and provide any text seed. The same names, team count, and seed always produce the same assignment, which makes the result easy to audit, share, or rerun. Every person is assigned once, team sizes differ by no more than one, and the request is rejected when there are more teams than people.

Prepare the list and choose a useful seed

Start with one entry for every person who should participate. Names are preserved exactly, including capitalization and spaces, so you can use display names, initials, employee labels, or other human-readable identifiers. Each entry must contain non-whitespace text. Repeated entries are allowed because two people can legitimately share a name, although adding an identifier is wise when the final assignment must distinguish them. Next, choose a team count from one up to the number of supplied names. The seed can be any non-empty text, such as an event name, a date chosen by your organization, or a published draw label. A descriptive seed makes the assignment easier to reproduce without keeping extra state. For a fair process, agree on or publish the seed before examining the teams. Changing even one character in the seed can change the shuffle, while using exactly the same input recreates exactly the same result. The tool accepts up to ten thousand names in one request, making it suitable for classrooms, workshops, leagues, and large internal events.

How deterministic balancing works

The generator converts the seed into a stable numeric state and uses that state to drive a deterministic shuffle of a copied name list. It does not use system randomness, the current time, network data, or stored session state. After shuffling, it deals names to teams in sequence: the first shuffled name goes to team one, the next to team two, and so on, returning to team one after the last team. This distribution rule guarantees that the largest and smallest teams differ by at most one person. When the name count does not divide evenly, the earlier numbered teams receive one additional member. Every input position appears exactly once in the output, so repeated display names remain separate entries rather than being silently merged. Determinism means reproducibility, not cryptographic unpredictability: anyone who knows the complete input and seed can calculate the same assignment. If participants might strategically influence a public seed, choose it through an agreed process after the final list is locked, then record both the list and seed alongside the result.

Rerun, compare, and automate assignments

Save the original order of the names, the requested team count, and the seed whenever you need to recreate an assignment. All three are part of the calculation: reordering the input list, correcting a name, adding a late participant, removing someone, changing the number of teams, or editing the seed creates a different shuffle. That sensitivity is useful because it prevents a stale result from appearing valid after the roster changes, but it also means you should treat the complete request as the record of the draw. The response returns the seed, the team count, and numbered teams with their members, which is straightforward to copy into a schedule or consume in an automated workflow. Use the browser runner for an occasional workshop or class, and use the API when an application needs repeatable grouping without maintaining its own randomization code. API requests cost $0.002; browser execution is available on the page. Validation stops impossible or ambiguous requests early, including empty names, a missing seed, a non-integer team count, or a team count greater than the number of people.