How long does the rate limit last on Twitter?

X does not publish a duration. The help page says the daily post limit is broken into smaller limits for semi-hourly intervals, so a post or reply limit belongs to one of those intervals and a daily limit to the day. The follow error tells you to try again in an hour or so. If the account is also asking you to verify a phone number or an email address, that is a lock rather than a rate limit, and waiting will not clear it.

Why does Twitter say I am rate limited?

Because one of the per-account limits X enforces has been reached: too many posts or replies in the current interval, too many follows in a day, too many direct messages, or too many email changes in an hour. The limits count every device and app on the account together, so a phone session and a browser session share the same bucket. X also says limits may be temporarily reduced during periods of heavy site usage, which is why the message sometimes appears at lower volumes than usual.

How do you fix rate limit exceeded?

You wait. The message means a limit bucket is full, and the bucket empties on its own when its interval elapses; nothing you do on your side speeds that up. What you can fix is the cause. Stop the action that filled it, check whether another device or a scheduled tool is spending the same account’s allowance, and lower your pace so it does not refill immediately. A script that retries on every rate limit error is itself the burst that keeps the limit engaged.

Is there a like limit on X?

Not a published one. The X limits help page lists daily figures for posts, replies, direct messages, follows and email changes, and says nothing about likes. That does not mean likes are unlimited. The automation rules say you may not like posts in an automated manner, and read-only mode, an enforcement action, removes the ability to like altogether. As a working figure, flottra caps likes at 12 per hour per account by default, and most days an account uses far fewer.

How many accounts can I follow per day on X?

The help pages say every X account can follow up to 400 accounts per day, and that once an account follows 5,000 accounts, further follows are limited by a ratio of followers to following that is calculated per account. The limits page calls 400 a technical account limit only and says there are additional rules prohibiting aggressive following, so 400 is not a safe target. A cautious automated account runs a fraction of that, spread unevenly across waking hours, and slower still while the account is new.

Does X Premium raise the rate limits?

The limits help page lists 50 original posts and 200 replies per day for unverified accounts, which implies verified accounts may be treated differently, but it does not publish the figure for them, and it publishes no Premium multiplier for follows, direct messages or anything else. Treat any specific number you see quoted for Premium accounts as unverified. For an operator the question is mostly moot, because the safe pace for an automated account sits well below the unverified figure anyway.

X rate limits in 2026: the official numbers and the real ones

X publishes daily limits per account: 50 original posts and 200 replies for unverified accounts, 400 follows, 500 direct messages, and a 5,000 following wall. No like limit is published. Rate limit exceeded means one of these buckets is full and clears on its own. The safe ceiling for automation is a fraction of the published one.

19 September 20268 min read

"Rate limit exceeded" is the most alarming message on X that means nothing is broken. A counter has hit the top of its bucket and will come down on its own. What matters is which counter, and why the published ceiling is not a safe one.

The first half of this page is what X (formerly Twitter, and still Twitter to most search engines) publishes, quoted as the help pages read on 19 September 2026. The second half is the operator's view: the much lower figures that keep an automated account alive.

The official limits, per account

X keeps its limits on one help page, About X limits, with the follow rules on a second, About following.

Action	Published limit	What the page adds
Original posts	50 per day	For unverified accounts; the daily limit is split into smaller limits for semi-hourly intervals
Replies	200 per day	For unverified accounts; the same semi-hourly split
Direct messages sent	500 per day	Counted across every device and app on the account
Follows	400 per day	Called "a technical account limit only"; aggressive following is prohibited separately
Following, total	5,000, then a ratio	Further follows are limited by account-specific ratios of followers to following
Account email changes	4 per hour	Stated per hour rather than per day
Likes	Not published	The page states no like limit
Replies per hour	Not published	Only the daily figure and the semi-hourly split are stated
Premium or verified accounts	Not published	The post figures are for unverified accounts; no multiplier is given

From About X limits and About following on help.x.com, checked 19 September 2026.

Posts and replies

The post limit, quoted exactly: "Posts: 50 original posts and 200 replies per day for unverified accounts. The daily update limit is further broken down into smaller limits for semi-hourly intervals." That last sentence explains most sightings of the message: you do not need to reach 50 posts, only the current half hour's share of them, and the page does not say how large that share is.

The page contradicts itself

Lower down, the same page still says: "The post limit of 2,400 updates per day is further broken down into semi-hourly intervals." The 50 and 200 figures in the current-limits list are the ones to plan around; a 2,400 quoted elsewhere came from that paragraph.

Two other lines matter. "These limits include actions from all devices, including web, mobile, phone, API, etc." And: "These limits may be temporarily reduced during periods of heavy site usage." So there is one bucket per account however you reach it, and the message sometimes appears at volumes that were fine last week.

Follows

The follow page adds the wall: "Every X account is able to follow up to 400 accounts per day." And: "Every X account can follow up to 5,000 accounts." Past that, further follows depend on a ratio of followers to following that is calculated for each account and not published, so nobody outside X can tell you where your wall is.

What "rate limit exceeded" actually means

The message means one of two things. Either a bucket belonging to one of the limits above is full for its interval, or X has reduced limits across the board, as it did with read limits in 2023 and as the help page allows "during periods of heavy site usage". Outside such periods, the last thing the account did is usually the bucket in question.

In both cases it clears on its own when the interval elapses. Logging out, reinstalling the app, switching device and clearing cookies all do nothing, because the count is kept on X's side against the account. Changing network is worse than useless: for an account that lives behind one proxy, a new address mid-session is a fresh inconsistency at a bad moment.

How long depends on the bucket, and X does not publish a duration. A post or reply limit belongs to one of the semi-hourly intervals. The follow error says "Try again in an hour or so." A daily limit clears with the day. Anything more precise is somebody's observation, not X's statement.

What it is not

Two other states get reported as rate limits and are not. The first is read-only mode, an enforcement action rather than a counter, described on the enforcement options page: the account can read its timelines and message its followers, but not post, repost or like. If that has outlasted any interval, it is not a rate limit.

The second is a lock. If the account asks you to verify a phone number or email address, X is questioning the account itself rather than counting its actions, and waiting will not clear it. See X account locked: how to unlock it.

The technical limit is not the safe limit

One line on the limits page applies to every row above: "The technical follow limit is 400 per day. Please note that this is a technical account limit only, and there are additional rules prohibiting aggressive following behavior." The number is what the interface allows. The rules about how you do it are enforced separately and come with no number.

The follow page names four prohibitions: "follow churn", "indiscriminate following" of unrelated accounts in a short time period "particularly by automated means", duplicating another account's followers, and services that claim to add followers. The automation rules put it in one line: "You may not follow or unfollow X accounts in a bulk, aggressive, or indiscriminate manner. Aggressive following is a violation of the X Rules."

The follow page gives four reasons for "You are unable to follow more people at this time": the daily limit; too many follows too quickly; the follow ratio; or a lock or limit "including due to aggressive follow behavior". The first two clear with time, the third moves only when your follower count does, and the fourth is enforcement wearing the same symptom.

An account that actually does 400 follows a day, every day, is describing a machine, and the rules against bulk and indiscriminate behaviour are what end it. X does not publish those thresholds. Assume the real ceiling for an automated account is a fraction of the technical one, and smaller still while the account is new; why X accounts get suspended covers what crossing it looks like.

What a cautious operator runs

The table puts the published ceilings next to the per account defaults flottra ships with, hourly caps and daily targets. They are ceilings rather than targets, and the gap between the columns is deliberate.

Action	Official limit (help page)	What a cautious operator runs
Original posts	50 per day, in semi-hourly buckets	3 per hour cap; four to six a day in practice, inside a daytime window
Replies	200 per day, in semi-hourly buckets	5 per hour cap; fewer on a new account
Follows	400 per day, "a technical account limit only"	10 per hour cap; 80 per day target, sized to what the machine can execute
Unfollows	No separate figure; follow churn is prohibited	50 per day; only after a 72 hour follow-back check, and 15 percent are kept anyway
Following, total	5,000, then a per-account ratio	Ratio brake at 1.05 once past 4,500 following, so the wall is never what stops it
Likes	Not published	12 per hour cap; most days far fewer
Direct messages	500 sent per day	3 per hour cap, inside a per-account window, under a campaign daily limit

Defaults per account. Every figure in the right column is a ceiling the account rarely touches, and the daily totals vary from day to day.

Two rows deserve a word. The follow target of 80 is a fifth of the technical figure, and a 15 percent chance of a rest day plus daily totals that vary by up to 35 percent mean the actual count differs every day. The unfollow row exists because near the 5,000 wall it is unfollow throughput that limits a follow-based routine, and fast unfollowing is what the follow churn rule describes. Both are covered in X follow bots and follow limits.

Posts go wrong in the other direction. An account posting anywhere near 50 a day reads as a feed rather than a person; four to six inside waking hours is what the default produces, and even that is a lot for a new account. Direct messages have their own guide, because 500 a day is the least interesting constraint on them: who you can message at all is the real one. Likes have theirs: no published limit is not permission.

Shape matters more than totals

Two accounts can make the same number of actions in a day and look nothing alike. The limits count totals per interval; the behavioural rules, and whatever detection sits behind them, see the shape: timing, gaps and repetition.

Even spacing is the most common tell. Follows at exactly the same interval, however long, are a metronome, and no person is a metronome, whatever the hourly total. Bursts are the second: nothing for hours, then a cluster inside a few minutes, is what a script that has just started looks like. Both fit inside every limit on this page, which is the point.

Quiet hours are the third. An account active at a steady rate around the clock is a machine, and one whose profile says London and whose activity peaks at 03:00 London time is contradicting itself. The window should follow the account's claimed location, which in practice means the proxy's. flottra's growth routine runs from 09:00 to 22:00 on the account's own clock by default, and that clock follows the proxy rather than the computer running the browser; choosing proxies for X accounts covers why.

Every device counts

The buckets include every device. If you touch an account by hand on your phone while a tool runs it in a browser, both spend the same allowance and shape the same pattern. An account that hits a limit at half its configured cap has found the other half somewhere you were not counting.

API rate limits are a different thing

If a developer tool reported that it was rate limited, that is a separate system. The X API has its own limits, set per endpoint and per 15-minute window and tied to the project's access tier. They govern how many requests an application may make, not how many posts an account may publish. The figures change with the tiers and are not reproduced here. What X automation can actually do covers the tiers and why they are not the route for running many accounts.

What to do when an account hits one

For one account that has just shown the message, stop and wait. For an operator running many, a rate limit is information worth reading first.

Identify the bucket. Posting, replying, following and messaging have separate limits, and the last thing the account did is the usual answer.
Stop everything on that account, not only the action that tripped it: a phone session, a scheduler nobody turned off, or a second tool on the same account are all spending the same allowance.
Do not retry in a loop. A script that retries every few seconds is the burst the behavioural rules describe. Wait out the interval, resume at a lower pace, and leave the cap lowered for days, not hours.
Check what the limit is hiding. If it arrives with a request to verify a phone number or email, or with posts that no longer appear in search, the account has moved into enforcement: see the locked account guide and the shadowban guide.

The rate limits are the one part of this X publishes, and the least important. They tell you where the interface stops; what decides whether an account survives happens below that line, in pace, shape and hours. Outside X's own heavy-load periods, an account that sees the message is telling you its pace was wrong.

Questions

How do I remove the Twitter rate limit?: You cannot remove it, and there is no setting, app or support request that lifts it early. A rate limit on X is a bucket of actions per interval that fills and then empties on its own once the interval elapses. Logging out, reinstalling the app or switching devices does nothing, because the limits count actions from all devices together. The only thing that shortens the wait is not hitting the limit again the moment it clears.
How long does the rate limit last on Twitter?: X does not publish a duration. The help page says the daily post limit is broken into smaller limits for semi-hourly intervals, so a post or reply limit belongs to one of those intervals and a daily limit to the day. The follow error tells you to try again in an hour or so. If the account is also asking you to verify a phone number or an email address, that is a lock rather than a rate limit, and waiting will not clear it.
Why does Twitter say I am rate limited?: Because one of the per-account limits X enforces has been reached: too many posts or replies in the current interval, too many follows in a day, too many direct messages, or too many email changes in an hour. The limits count every device and app on the account together, so a phone session and a browser session share the same bucket. X also says limits may be temporarily reduced during periods of heavy site usage, which is why the message sometimes appears at lower volumes than usual.
How do you fix rate limit exceeded?: You wait. The message means a limit bucket is full, and the bucket empties on its own when its interval elapses; nothing you do on your side speeds that up. What you can fix is the cause. Stop the action that filled it, check whether another device or a scheduled tool is spending the same account’s allowance, and lower your pace so it does not refill immediately. A script that retries on every rate limit error is itself the burst that keeps the limit engaged.
Is there a like limit on X?: Not a published one. The X limits help page lists daily figures for posts, replies, direct messages, follows and email changes, and says nothing about likes. That does not mean likes are unlimited. The automation rules say you may not like posts in an automated manner, and read-only mode, an enforcement action, removes the ability to like altogether. As a working figure, flottra caps likes at 12 per hour per account by default, and most days an account uses far fewer.
How many accounts can I follow per day on X?: The help pages say every X account can follow up to 400 accounts per day, and that once an account follows 5,000 accounts, further follows are limited by a ratio of followers to following that is calculated per account. The limits page calls 400 a technical account limit only and says there are additional rules prohibiting aggressive following, so 400 is not a safe target. A cautious automated account runs a fraction of that, spread unevenly across waking hours, and slower still while the account is new.
Does X Premium raise the rate limits?: The limits help page lists 50 original posts and 200 replies per day for unverified accounts, which implies verified accounts may be treated differently, but it does not publish the figure for them, and it publishes no Premium multiplier for follows, direct messages or anything else. Treat any specific number you see quoted for Premium accounts as unverified. For an operator the question is mostly moot, because the safe pace for an automated account sits well below the unverified figure anyway.