Instagram Says 100 Posts per 24 Hours. The Same Page Also Says 50.

If you have been trying to find Instagram's daily posting limit and getting contradictory answers, here is why: the contradiction is in Instagram's own documentation.

Both of these are on Meta's Instagram Platform content publishing page, checked 2026-09-15:

"Instagram accounts are limited to 100 API-published posts within a 24-hour moving period."

"Accounts are limited to 50 published posts within a 24-hour period. Publishing a carousel counts as a single post."

The second appears in the carousel section. The wording is nearly identical to the first. The number is half.

This is a different shape from the usual documentation conflict

We have catalogued several kinds of conflicting platform numbers, and they are not all the same problem:

The last row is the rarest and, oddly, the most reassuring. It is not evidence that nobody knows the limit. It is evidence that a document got edited in one place and not another. Documentation drift is a much more boring explanation than a secret rule.

Which one to plan against

Fifty.

The rule we apply whenever two official numbers disagree is to take the conservative one, and the reasoning is asymmetric: if you plan against 50 and the real limit is 100, you lose a little headroom you were not using. If you plan against 100 and the real limit is 50, you find out at the moment your queue backs up — which will be the day you most needed it not to.

When two official numbers disagree, the cost of being wrong is not symmetric. Plan against the one where being wrong is cheap.

What the numbers do and do not cover

Read the qualifiers carefully, because they rule out most of the people worrying about this:

• Both say "API-published" / "published posts" in a publishing-API context. They govern posts made through the platform API.
• Posting manually from the app is not counted by these numbers.
• And Instagram has never published a manual daily posting cap. Given a document that specifies "carousels count as a single post" and "carousels are limited to 10 images, videos, or a mix of the two", a missing manual cap is very unlikely to be an oversight.

So if your question was "how many times can I post from my phone today", the answer is that there is no published number, and the thing people experience as a limit is usually a distribution effect: several posts in quick succession compete with each other, and the later ones get less independent reach. That is not a penalty and there is nothing to appeal.

The better question: what does your account say?

Meta documents an endpoint whose entire job is to report the current usage for a specific account — referenced in the docs as content_publishing_limit. It tells you how much of the rolling window that account has consumed right now.

This is more useful than any static number for two reasons: it is your account rather than an average, and it is current rather than whenever the doc was last edited.

If you use a third-party scheduler, it is also an excellent vendor question: "do you read the publishing limit endpoint?" A tool that reads it can slow down before hitting the wall. A tool that does not can only hit the wall and report an error.

If you are anywhere near this limit

Then you are almost certainly not running one account, and the interesting constraints are elsewhere: whether each account has its own environment, whether the content is genuinely different or one piece fanned out, and whether the accounts publish in the same second.

Those three decide outcomes far more than a 24-hour quota does — which is the problem NoobClaw addresses, with per-account personas producing different content, randomised pacing, and publishing through your own logged-in sessions in isolated local browser profiles rather than one shared API token. For a single account, none of this is relevant and neither limit will ever bind you.

A general method for platform numbers that disagree

This case is a good template, because the same situation comes up constantly and people resolve it badly — usually by picking whichever number the first search result gave them.

Work through these four questions in order:

1. Is it in official documentation at all? If the only sources are third-party articles, treat the number as folklore until proven otherwise. Numbers that circulate in five different versions are usually numbers that do not exist.
2. Are the two numbers describing different tiers? Documentation frequently lists limits by device, plan or surface in one table. That is not a contradiction, it is a tier list, and your job is to find your row.
3. Are they on different product surfaces? Two official pages can legitimately disagree because they describe two different pipelines maintained by two different teams. That is the LinkedIn member-versus-Page pattern.
4. Same page, two numbers? Then it is documentation drift. Plan against the conservative one and move on.

The order matters. Skipping to step four makes you read a perfectly clear tier table as a contradiction, which is how people end up planning around a number that was never meant for them.

And whenever the platform offers a way to query your own current state, that beats all four steps. A number in documentation describes a policy; an endpoint describes your account. Where both exist, the endpoint wins.

We have run this method on several other numbers people search for constantly — the Close Friends limit, which turned out to originate in an error message rather than a specification, and the daily unfollow limit, where every circulating version has a different number and none has a source. The pattern repeats often enough to be worth internalising once.

FAQ

So is it 100 or 50?

Both sentences are currently in the official documentation. Plan against 50, and if your volume is close to either number, read your real usage from the publishing limit endpoint rather than trusting a static figure.

Do Reels and carousels count separately?

Carousels explicitly count as one post regardless of how many items they contain (maximum 10). The documentation does not split Reels into a separate quota, so the safer assumption is that they share it.

Does this limit apply to Creator Studio or Meta Business Suite?

These figures appear in the developer documentation for the publishing API used by third-party applications. First-party surfaces are not third-party API consumers in the same sense — if this matters to your workflow, verify it in your own account rather than inferring it. The first-party versus third-party trade-off more generally is in Meta Business Suite vs Buffer, and the earlier single-number version of this article is here.