How many posts per day each platform allows, and what works

Short answer. Of the ten platforms below, six publish no posting limit at all, and where a number does exist it is 20 to 10,000 times higher than any cadence a practitioner recommends. Only Bluesky publishes real per-hour numbers (1,666 record creations an hour); Instagram and Threads publish API publishing caps (100 and 250 per 24 hours); X publishes 50 original posts a day for unverified accounts. TikTok, YouTube, LinkedIn, Facebook Pages, Reddit and Pinterest publish nothing you can plan against. All retrieved 2026-09-21.

Which means the question "how many posts per day am I allowed" has almost no useful answer, and the question underneath it, "how many should I post", has a completely different one.

6 of 10

platforms that publish no posting limit

retrieved 2026-09-21

50 / day

X original posts, unverified accounts

help.x.com, X limits

1,666 / hour

most records Bluesky lets you create

bsky.network, rate limits

2-3 / day

X originals practitioners recommend

practitioner consensus

The table

The hard-limit column is the platform's own documentation, with the URL and the date we read it. Where it says not published, we looked and there is nothing; we are not filling the cell with a number from a vendor blog. The cadence column is practitioner consensus, which is a weaker kind of claim and is labelled as one everywhere it appears.

Bluesky's documentation is the only one written like an engineer wrote it for another engineer, and it is worth reading the sentence it ends on: "These limits shouldn't affect typical Bluesky users! They won't affect the majority of developers either. But they will affect prolific bots, such as the ones that follow every other account or like every post on the network." That is what every one of these numbers is for. None of them is a cadence recommendation.

What "not published" hides

Six rows in that table are blank on purpose, and each one is blank for its own reason.

LinkedIn is the most explicit about being unhelpful. Its invitation-limit page says a restriction "typically lasts one week", that "withdrawing pending invitations will not remove the restriction", and that "LinkedIn Support cannot disclose the type or reason for the restriction." The number everyone quotes, 100 invitations a week rising to 200, appears nowhere in LinkedIn's own material. The claim that a high Social Selling Index raises that ceiling is also unsourced, which we go into in the SSI note.

YouTube confirms the cap and withholds the number in the same paragraph: "we limit how many videos a channel can upload in a 24-hour period across desktop, mobile, and YouTube API. Limits may vary by country/region or channel history." So the limit is not a constant, it is a function of your standing. That sentence is more informative than a number would be.

TikTok publishes API request rates and nothing about how many videos a human may post. Facebook publishes an app-level call budget that scales with your app's own daily active users, which is a developer quota and not a page posting limit. Reddit publishes a query-per-minute figure for data access and nothing about submissions, because on Reddit the constraint is per-subreddit rules and the spam filter, not an account cap. Pinterest publishes request quotas only.

The pattern is consistent and worth naming: platforms publish limits for machines and withhold limits for people. Anything that reads like a posting cap for a human is either from a developer doc or from a vendor blog guessing.

Why "more posts" backfires on a new account

The hard limits are so far above any sensible cadence that they are never your constraint. Something else is, and it is the same mechanism on every ranked feed.

Each post is tested on a small cohort first. New accounts have no interest signal, so the first posts are tested against a generic audience and underperform until the recommender works out who you are for. That test audience is a fixed, small pool. Post twice in a day and your second post is tested on roughly the people your first one just reached. You did not double your at-bats; you split them. On LinkedIn this is explicit in the product: an author-diversity reranker scales down a second post from the same author inside one viewer session, so the second post of the day is mechanically handicapped, not just unlucky. On TikTok the equivalent is splitting the seed cohort that your best video needed whole.

Repetition trips a different system entirely. Recycled, near-duplicate and low-substance posts are now demoted by name. LinkedIn's March 2026 feed announcement says it is "reducing recycled and click-driven posts"; the originality rules on YouTube, TikTok and Instagram operate at the account level rather than per post, which is the part people miss, and which we cover in unoriginal content rules. Volume is the fastest way to run out of genuinely original things to say, and the penalty lands on the account, not on the one weak post.

Zero views on a new account is usually the cold-start test, not a ban. If that is why you are here, the distinction matters a lot and we take it apart in TikTok 0 views on a new account and, more generally, in what is a shadowban.

The two readers of this page

Almost everything written about posting limits is written for one reader and read by another. It is worth separating them out loud.

Reader A: a founder with one real account. Every number in the table above is irrelevant to you. You will never send 50 posts in a day. Your constraints are attention, originality and the cold-start test. The right move is the low end of the cadence column and consistency inside one topic, and the only limit on this page you might realistically meet is LinkedIn's undocumented invitation ceiling.

Reader B: an operator running many accounts. The hard limits are your operating envelope, and they are per account, so they do not aggregate against you; the app-level budgets do. X's 10,000 posts per 24 hours is per app, not per account; Facebook's call budget scales with your app's users; Bluesky's relay caps new PDS instances at 100 accounts by default. That last one is the tell: Bluesky sized its defaults around the difference between a person and a fleet, and wrote it down.

We are not reader B, and we will say plainly that the well-known advice on this topic, residential proxies, per-account browser fingerprints, two-hour action spacing, drip schedules, is reader-B advice. For a founder with one genuine account it is inert at best. The one part of it that actively hurts reader A is the automation software itself: LinkedIn says it takes action against "unauthorized third-party tools", and that is a real risk you take on for a benefit you do not need. What a warmed account actually costs, and who buys them, is its own subject: what a warmed TikTok account costs.

A starting week

This is the cadence we would run on a brand-new account for the first month. It is the low end of every practitioner band, because a new account's problem is never volume.

Two things to notice about that table. Reddit's week-one row has no posts in it, because on Reddit the account has to earn the right to post before posting is useful, which we go through in Reddit karma and CQS. And nothing in the table is near a published limit, anywhere, by a factor of at least twenty.

About the dates

Every number here carries the date we read the page, because these move. X changed its own post cap and left a contradicting sentence further down the same help page, which we untangle in X rate limit exceeded. Instagram's publishing cap and Threads' 250 are the kind of figure that gets raised quietly. We re-check this table because our own publishing engine breaks when these numbers move, which is a better incentive than an editorial calendar. If a cell here is stale, the date in the header is the first thing that will tell you.

If a platform has shown you an actual restriction rather than a rate limit, that is a different diagnosis: start at account restricted: what platforms tell you.