Starlight Tools

Random Picker — Pick a Random Item or Winner from a List

Pick names or choices, draw multiple winners with or without repeats, and keep your list private—no signup required.

Add your list

Try an example:

Auto-detects newlines, commas, semicolons, and tabs. Tested with up to 10,000 entries.

0 options · 0 unique

Draw settings

Maximum updates with the available pool; with replacement is capped at 1,000.

Selected result

Your random choice will appear here.
Random source will be checked when you draw.

List tools and more options

Open list tools, import, save, share, weights, and seeded mode

Import and cleanup

Pastes and files are processed in this browser; they are not uploaded.

Weighted choices

Enter Pizza | 3 and Sushi | 1, or paste two tab-separated columns. Weight must be greater than zero.

Saved lists

Presets use localStorage on this device. They are optional and never sent to us.

Share this list

Verifiable seeded mode

Seeded mode uses deterministic ST-RP-v1, records a SHA-256 list hash, and is reproducible rather than unpredictable. It is not third-party certification.

Draw history (this tab only)

No draws yet.

How to use the random picker

  1. Type, paste, or upload choices. Auto format recognizes newlines, commas, semicolons, and spreadsheet tabs.
  2. Choose the winner count and turn on No repeats when every winner must be unique.
  3. Select Pick Random. Results, the remaining pool, and a session-only history appear immediately.
  4. Use Remove winner and pick again for elimination-style rounds, or open List tools for weights, presets, sharing, shuffling, and seeded receipts.

Lists, uploaded files, ordinary results, and history stay in this browser. A saved preset is local to this device. A share link exposes its encoded contents to anyone who receives it and is created only after explicit consent.

Practical examples

Choose dinner

List: Pizza, curry, tacos, ramen

Settings: 1 winner; repeats either way.

Illustrative result: Tacos.

Pick a classroom student

List: Ava, Ben, Chloe, Diego, Emi

Settings: 1 winner; No repeats on for a rotation.

Illustrative result: Chloe.

Select three giveaway winners

List: One entrant per line.

Settings: 3 winners; No repeats on.

Illustrative result: 1. Noor, 2. Sam, 3. Taylor.

Assign a speaking order

List: Team North, Team East, Team South, Team West

Settings: Open List tools and choose Shuffle all.

Illustrative result: East, West, North, South.

Allow repeat selections

List: Clean board, answer phone, greet visitors

Settings: 4 winners; No repeats off.

Illustrative result: Greet visitors may appear more than once.

Fairness, probability, and algorithm

For N unique equal-weight options: P(item) = 1/N. If an entry appears k times among T parsed entries: P(entry) = k/T. When selecting m unique winners without replacement from N unique entries: P(entry is included) = m/N.

Worked example: With five unique names—Ava, Ben, Chloe, Diego, and Emi—each has a 1/5 = 20% chance on a single equal draw. If two winners are selected without replacement, each name has a 2/5 = 40% chance of appearing among the two winners.

Equal mode trims entries and ignores empty values. Duplicates remain separate tickets unless Remove duplicates is used. Weighted mode parses a positive weight after | or in a second tab-separated column and shows each entry's probability before drawing. No-repeat weighted draws remove the selected label and its weight before the next selection.

Fairness and algorithm details

Algorithm version: ST-RP-v1. Ordinary draws request unsigned 32-bit values from crypto.getRandomValues. Rejection sampling discards values outside the largest multiple of the requested range, avoiding modulo bias. If the browser API is unavailable, the page visibly reports a Math.random fallback and does not describe that draw as secure.

Shuffle all uses Fisher–Yates, selecting each swap index with the same active random source. Seeded mode derives a deterministic PRNG state from the public UTF-8 seed, records a SHA-256 hash of the normalized list, and produces a receipt containing the settings and result. A seeded receipt improves reproducibility but does not prove that the published input was complete or certify the organizer.

Limitations: A private browser draw is not independently verifiable, audited, or certified. crypto.getRandomValues provides a strong browser randomness source, but the overall giveaway procedure still requires appropriate rules, records, and oversight. Last browser-logic test: July 17, 2026. Documented list limit: 10,000 parsed entries.

Frequently asked questions

How do I pick multiple unique winners?

Set Winners to the number needed and select No repeats. The picker draws without replacement and will not allow a winner count above the number of unique entries.

What do duplicate entries do to the odds?

In equal mode, duplicates act like extra tickets. If an entry appears k times among T entries, its probability on one draw is k/T. Remove duplicates when every unique entry should have the same chance.

What is the difference between replacement and no replacement?

With replacement, every draw uses the full pool and the same entry can win again. No repeats draws without replacement, removing each selected entry from that draw's pool.

Do my entries leave this device?

No. Picking, uploads, saved presets, and history are processed in your browser. A share link contains the list in its URL only after you explicitly consent, and anyone with that link can read it.

How large can my list be?

The documented tested limit is 10,000 parsed entries. The page warns above 5,000 entries, blocks larger lists, and caps a with-replacement draw at 1,000 winners to keep browsers responsive.

Can I paste a spreadsheet or upload CSV and TXT files?

Yes. Auto format accepts newlines, commas, semicolons, and tabs, including Excel and Google Sheets pastes. TXT and CSV files are read locally in the browser and are not uploaded.

How do I remove a selected winner?

After a draw, choose Remove winner and pick again. Every source entry matching the current winner or winners is removed, then a new draw runs from the remaining list.

Can a result be reproduced?

Ordinary private draws cannot be reproduced. Advanced seeded mode records a public seed, SHA-256 normalized-list hash, settings, and algorithm version in a receipt so the same normalized input and settings reproduce the draw.

Is this suitable for giveaways?

It is useful for informal giveaways, especially with saved history or a seeded receipt, but a private browser draw is not independently audited or certified. Follow applicable rules and use an appropriate third-party audited process for high-trust promotions.

Is every list order equally likely after shuffling?

Yes, assuming the active random source behaves uniformly. The page uses Fisher–Yates and chooses each swap index uniformly with rejection sampling when secure browser randomness is available; seeded mode uses its documented deterministic generator.

Explore more tools