Draw winners with a public seed and a receipt anyone can check
In a giveaway, the result is only worth as much as the audience's trust in it. So this tool does not simply pick a name: it makes the result reproducible. You paste the participant list, set the prizes and enter (or generate) a public seed. The draw is a deterministic function of the seed and the list, so anyone holding the receipt and the list lands on exactly the same winners, as many times as they want. It also reads weights ("Ana, 3" is worth three chances), takes CSV files, shows which names are repeated, draws several named prizes with backups in queue order, and runs successive rounds excluding whoever already won. Everything happens in the browser, and the Verify tab replays the draw described in a receipt to tell you whether it holds up.
The draw starts from a seed, a short piece of text you choose or the tool generates. That seed is combined with the SHA-256 fingerprint of the participant list and feeds an xoshiro128** pseudorandom generator. Because the generator is deterministic, the same seed with the same list always produces the same winners, which is what lets anyone redo the calculation.
The order of the winners comes from a Fisher-Yates shuffle using integers drawn by rejection sampling, that is, without the modulo bias you get when someone simply takes the remainder of a division. With weights on, the draw becomes sequential sampling without replacement using a cumulative sum plus binary search: someone with 3 chances is exactly three times as likely as someone with 1, and leaves the pool once they win.
The receipt gathers the date, the seed, the SHA-256 fingerprint of the list, the prizes, the winners, the backups and who was kept out of the draw. It proves those winners are the result of that seed with that list: swap a name in the announcement and verification catches it, edit the list and the fingerprint stops matching.
What it does not prove on its own is that the list was that one before the draw. A dishonest organizer could run the draw many times with different seeds until they liked the outcome, and only then publish it. The defence is simple and it is yours: publish the seed first (in a story, a pinned comment, a message to the group) and draw afterwards. The seed then becomes a public commitment and the receipt closes the loop.
Comment giveaway: paste the participants' handles, turn on "merge repeated names" so that someone who commented ten times does not get ten chances (or leave it off if that is exactly your rule) and use the list of repeats to see who showed up more than once before drawing.
Raffle with sold numbers: paste one line per number, or use weights, writing "Ana, 12" for someone who bought twelve tickets. Named prizes cover the "first place takes the bike, second takes the headphones" case, and the backups already define who steps in if someone does not answer. For a second round, keep "exclude winners from previous rounds" on: the tool redraws with whoever is left and the receipt records who was kept out.
Paste the code into your HTML and the tool shows up on your page, without J-Kit's navigation and ads. It still runs in the browser of whoever visits your site.
<iframe
src="https://jkit.tools/embed/en-US/giveaway-winner-picker"
width="100%"
height="600"
style="border:0"
loading="lazy"
title="Giveaway Winner Picker"
></iframe>Publish the seed before drawing and the receipt afterwards. The receipt carries the seed, the SHA-256 fingerprint of the list and the winners. Anyone can paste those into the Verify tab and watch the same result appear, with no need to trust you or us.
generated at draw time
Everything runs in your browser. The list is saved only on your device and is never sent to servers.