Tool
Random Seed Generator
Generate a seed, then reproduce the exact same draw from it — the auditable kind of random.
Random Seed Generator
Generate a seed, run a draw from it, and get the same result every single time from that seed — on this device, on any device, today and next year.
That is the opposite of what every other tool on this site does, and it is deliberate. Reproducibility and unpredictability are mutually exclusive, and a draw somebody has to be able to check afterwards needs the first one.
The seed is a readable phrase rather than a raw integer, because a seed that has to survive being read aloud on a stream, typed into a comment, or written on a whiteboard needs to be something a human can carry.
How the Random Seed Generator works — and why it's fair
Your seed phrase is hashed into a 32-bit integer with FNV-1a, and that integer initialises a small pseudo-random generator called mulberry32. Given the same starting integer, mulberry32 produces exactly the same sequence of values forever — that is what a pseudo-random generator is, and it is why the first value from any given seed is shown on screen before you draw anything.
The shuffle itself is Fisher-Yates, identical to the unseeded order randomizer, with one substitution: every random index comes from the seeded generator rather than from crypto.getRandomValues. Same algorithm, same uniformity guarantee across orderings, entirely different property in time — the seeded version is a function of the seed, and the unseeded version is a function of nothing you can reproduce.
This is the single most misunderstood distinction in practical randomness. A pseudo-random generator is not a worse random generator; it is a deterministic function that produces well-distributed output. For a draw, that determinism is either the most valuable property available or a complete disqualification, depending entirely on whether anybody needs to verify the result later.
The verification workflow is straightforward and does not require anyone to trust this site: publish the entrant list and the seed, and any reader can reproduce the ordering here and confirm it matches what you announced. The stronger version of this is commit-and-reveal — publish a hash of the seed before the draw and the seed itself afterwards, which additionally proves you did not pick the seed after seeing the outcome.
The generated seeds are three words plus a number, drawn from a small word list. That gives a few thousand possibilities, which is plenty for readability and nowhere near enough for secrecy. A seed is meant to be published, so this is the right trade — but it does mean a seed is never a secret, and anyone who knows it knows every value it will produce.
When the Random Seed Generator is fair — and when it is not
What it does guarantee
- The same seed always produces the same ordering, on any device and at any time.
- The shuffle is Fisher-Yates, so for any given seed every ordering was equally reachable before the seed was chosen.
- Anyone can reproduce your draw from the published seed and entrant list, without trusting this page.
What it does not
- It is completely predictable to anyone who knows the seed. Never use it where the outcome must be secret.
- A seed chosen after seeing outcomes proves nothing. Commit to the seed before the draw, publicly.
- The generated seed phrases come from a small word list. They are readable identifiers, not secrets.
Two worked examples
Publishing a draw for a giveaway
- Post the frozen entrant list and the seed 'amber-flint-glade-472' before drawing.
- Run the seeded shuffle; the first name is your winner.
- Anyone can paste the same list with the same seed and get the identical order — no trust required.
The same seed on two devices
- Seed 'cedar-hollow-larch-108' hashes to the same 32-bit integer everywhere.
- mulberry32 from that integer produces the identical value sequence on both machines.
- So a five-name list shuffles into the same order on a phone in one country and a laptop in another.
Draws that have to survive a question
Public giveaways are the main case, and the value is entirely in what happens after the announcement. A winner picked by an unseeded draw is unverifiable by definition — nobody can check it, including you. A winner picked from a published seed can be re-derived by any sceptic in about ten seconds.
Research and analysis use it for a different reason: a randomised assignment that cannot be reproduced cannot be reviewed. Recording the seed alongside the results turns a random split into a documented one, which is the difference between a result and a finding.
Tournament and league organisers use it for fixture draws where the accusation of favouritism is routine. Publishing the seed in advance and drawing in public removes the accusation entirely rather than answering it.
For anything that must be unpredictable — a password, a PIN, a prize nobody should be able to predict — use the tools built on crypto.getRandomValues instead, and read the guide on how the two kinds of randomness differ before choosing.
Frequently asked questions
Is a seeded draw less random?
Its output is well-distributed and its orderings are uniform. What it is not is unpredictable — anyone with the seed can compute every result in advance.
When should I not use this?
Any time the outcome must be secret or unpredictable in advance. Passwords, PINs, and any draw where knowing the result early confers an advantage.
Why words instead of a number?
Because a seed has to survive being read aloud, typed by someone else and written on a whiteboard. 'amber-flint-glade-472' does that; 2947183624 does not.
Can someone predict my draw from the seed?
Completely, and that is the intended property. A seed is something you publish, not something you protect.
How do I prove I did not pick the seed after the draw?
Publish a hash of the seed before drawing and the seed itself afterwards. That is commit-and-reveal, and it is covered in its own guide.
Does the same seed work on other people's devices?
Yes. The hash and the generator are both deterministic and platform-independent, so the same seed and list give the same order everywhere.
How many seeds can the generator produce?
A few thousand from the built-in word list. That is ample for readability and far too few for secrecy, which is fine because seeds are meant to be public.
Can I use my own seed phrase?
Yes, type anything. A memorable phrase, a date, a public event — anything that hashes to an integer, which is everything.
Is this the same generator the other tools use?
No. Every other tool here uses crypto.getRandomValues. This page is the only one that deliberately uses a pseudo-random generator, and it is separated for exactly that reason.
Related tools
Order Randomizer
Shuffle any list into a fair random order — presentation order, turn order, anything.
Raffle & Giveaway Picker
Paste your entrant list and draw a fair, verifiable winner — built for streamers.
Random Sample From a List
Draw a fixed-size sample without replacement — for audits, spot checks, and surveys.