You ran the giveaway, the comments rolled in, and now you owe someone a prize. Picking the winner is the part most people wing, and it's the part that decides whether your followers trust the next one. This guide walks through every way to do it, when each one stops working, and how to end up with a winner you can prove was chosen at random.

Key takeaways

Picking by hand is fine for a few dozen comments. Past roughly 50, it becomes slow, error-prone, and impossible to prove.
A good picker does three things: reads every public comment, applies your rules consistently (one entry per person, tagged friends, keywords), and draws with randomness you can verify afterward.
Pickfetti does all of that from a public post link, with no Instagram login, and gives you a shareable proof page with the seed and hash. Free for posts up to 500 comments.
Always draw alternates at the same time as your winners so a no-show doesn't force a second, less trustworthy draw.
Publish the proof link with your announcement. Proof beats "trust me."

The four ways people pick a winner

1. By hand (scroll and point)

Open the post, scroll, close your eyes, tap. It's honest in spirit, but there's no way to show anyone that you didn't just pick your friend, and Instagram's comment ordering isn't random, so the "top" comments get picked far more often. Fine for a 20-comment giveaway among people who already trust you. Not fine for anything you'd call a promotion.

2. Screenshots or a screen recording

Some hosts record themselves scrolling and stopping. It feels transparent, but a recording only proves that a comment existed and that you stopped on it. It doesn't prove the stopping point was random, that every comment had an equal chance, or that duplicates were removed. See what makes a draw provably fair for why recordings alone don't cut it.

3. Spreadsheet

Copy the usernames into a sheet, dedupe, number them, then use a random number generator. This is real work, but it does produce a defensible result if you keep the sheet. The pain is the copying: Instagram doesn't give you a "download comments" button, so you're pasting hundreds of names by hand and hoping you didn't skip a page.

4. A random comment picker

A picker reads the comments for you, applies your rules, and draws. The differences between tools come down to how they get the comments (login versus public URL), how they handle rules like "tag 2 friends," and whether they give you proof you can hand to your audience.

Why manual picking breaks past ~50 comments

Three things go wrong at the same time:

Duplicates. The same person comments five times ("tagging more friends for extra entries!"). If you don't dedupe, they get five tickets while everyone else gets one, and if you do dedupe by hand you'll miss some.
Rule checks. "Tag 2 friends" sounds simple until you're counting @-mentions in 400 comments and deciding whether tagging the same friend twice counts.
Provability. With no record of the entrant list and no record of the random choice, all anyone has is your word. That's usually enough right up until it isn't.

None of that is a criticism of anyone's honesty. It's just that "fair" and "provably fair" are different things, and the second one needs tooling.

How Pickfetti picks a winner

Paste the post or reel link. Any public Instagram post works. Pickfetti reads the public comments the way a logged-out visitor would see them. You never enter an Instagram password and never connect an account. (Private accounts can't be read; Stories can't be used because they have no public comments.)
Set your rules. Number of winners and alternates, one entry per person or every comment counts, minimum @-mentions, required keyword or hashtag or emoji, exclude the post author, a blocklist of specific accounts, ignore replies, a date/time window, and optionally "winners must follow @host."
Draw. Pickfetti generates a seed from the operating system's cryptographic random source, then runs a deterministic HMAC-SHA256 Fisher–Yates shuffle over the eligible entries. Same seed plus same entries always gives the same order, which is exactly what makes it checkable.
Share the proof. Every draw gets a public page at pickfetti.com/d/{code} showing the winners, the rules, the seed, and a hash of the entrant list. Anyone can request /d/{code}/verify.json and recompute the result themselves. You also get a winner card image for your announcement and a CSV export of entries.

The whole thing takes about the time it takes to read this paragraph. Try it on a post; there's no signup for posts up to 500 comments.

Handling "tag 2 friends" rules

This is the most common rule and the most commonly botched one. A few things worth deciding before you draw:

Do repeated tags count? They shouldn't. Pickfetti counts only different accounts, so "@bestie @bestie" is one tag, not two.
Do tags across several comments add up? Pickfetti pools a person's mentions across all their comments on the post. Someone who tagged one friend in each of two comments has tagged two friends. That matches how most people naturally read the rule.
Does tagging the host count? If you want, add the host to your blocklist of accounts that don't count, or just accept it. Be consistent.

Set the minimum mentions to 2 (or whatever your caption said) and the picker enforces it before the draw, so ineligible comments never enter the pool. If you're writing the caption now, the rules generator turns your settings into clear caption wording.

Duplicates and "extra entries"

Decide up front. Pickfetti's default is one entry per person, which is what most audiences expect and what "fair" usually means. If your caption promised "every comment is an entry," switch to every-comment mode so the count matches your promise. What you shouldn't do is promise one thing and draw the other, because a screenshot of your caption next to your proof page will show the mismatch.

Alternates (draw them now, not later)

Winners ghost. They change their username, they don't check requests, they turn out to be a bot. If you draw a fresh winner a week later, you've created a second draw that skeptics can question. Instead, draw one to three alternates in the same draw. Pickfetti lists them in order on the proof page, so "our first winner didn't respond in 48 hours, moving to alternate #1" is a completely transparent step. More on this in how to announce a winner.

Follow requirements

"Must follow us to win" is fine as a rule, but no public tool can check the follower list of every commenter, and you wouldn't want it to. Pickfetti checks the drawn winners only: if a winner provably doesn't follow the host, they're skipped and listed transparently, and the next person in the shuffle moves up. Private accounts show as "unknown" and are never rejected for it, because you can't punish someone for something you can't see. Likes and "shared to your story" simply can't be verified from public data, so don't hang eligibility on them.

A quick word on comment counts

If the picker retrieves fewer comments than the number Instagram shows on the post, that's normal: deleted, hidden, and spam-filtered comments still count in Instagram's tally, and Facebook cross-post comments aren't visible on Instagram at all. Pickfetti shows both the retrieved and reported numbers so you can see the gap. Here's the full explanation.

Checklist before you hit "draw"

Rules in the picker match the rules in your caption.
Post author excluded (and any team accounts on the blocklist).
Date window set if the giveaway had a stated close time.
Winners plus alternates chosen.
You're ready to paste the proof link into your announcement.

FAQ

Can I pick a winner without logging in to Instagram?

Yes. Pickfetti reads public comments from a post link. No login, no password, no account connection. It can't read private accounts, and it doesn't need to.

How many comments can I use for free?

Posts with up to 500 comments are free with unlimited draws, no account and no card. Bigger posts use a one-time Draw Pass ($4.99, up to 10,000 comments) or Big Draw Pass ($14.99, up to 100,000), or Pro at $9/month. See pricing.

Can people verify the winner really was random?

Yes. The proof page publishes the seed and the hash of the entrant list; the verify.json endpoint lets anyone recompute the shuffle and confirm the same winners come out.

What if the winner doesn't respond?

Move to the next alternate from the same draw. Announce that you're doing so and link the same proof page.

Can I pick from a reel?

Yes, any public post or reel link works. Stories can't be used because they don't have public comments.