"How many DMs can I send before Instagram blocks me?" is one of the most searched questions in this space, and almost every answer to it is somebody's guess presented as fact.
The confusion is that two completely different sets of limits get discussed as though they were one thing:
- Manual sending — you, in the app, typing messages. Undocumented, behavioural, and the source of every horror story.
- The messaging API — a tool sending on your behalf through Meta's official infrastructure. Documented, precise, and much more boring.
They behave differently enough that advice about one is actively misleading about the other.
Manual DM limits: what nobody actually knows
Meta does not publish a number. Anyone who gives you one — 50 a day, 80 a day, 150 a day — is repeating a figure someone else invented.
What is genuinely known is that Instagram's anti-spam systems exist, that they are behavioural rather than a simple counter, and that they respond to patterns rather than volume alone.
The patterns that reliably attract attention:
Identical messages to many people. The single strongest signal. Copy-pasted outreach to forty accounts reads as automation regardless of whether you typed it yourself.
Messaging people who have no relationship with you. DMs to accounts that do not follow you and have never interacted go to message requests, and a high volume of them from a new account is treated very differently from replies to your own commenters.
Speed. Thirty messages in four minutes is not a human pattern.
Reports. The fastest route to a restriction is people marking your messages as spam. Two or three reports does more damage than several hundred messages nobody minded.
Account age and history. A three-week-old account with 40 followers gets a fraction of the latitude of a two-year-old account with an established posting history.
The practical version: volume is not what gets accounts restricted — pattern is. A hundred genuine replies to people who messaged you first is unremarkable. Twenty identical pitches to strangers is not.
API limits: what is actually documented
This is the part with real numbers, and they are far more restrictive than most people assume.
| Limit | What it means |
|---|---|
| One private reply per comment | Exactly one automated DM per comment. Not a sequence. |
| Seven days | The private reply must be sent within seven days of the comment. |
| 24-hour standard messaging | After someone messages you, you have 24 hours to reply freely. |
| No cold messaging | There is no endpoint for DMing someone who never interacted with you. |
| Professional account required | Business or Creator. Personal accounts have no API access. |
| Rate limits | Published caps on API calls per hour, enforced by Meta. |
Notice what is missing: there is no daily DM cap in the API. A tool sending private replies to 4,000 commenters in a day is not violating a volume limit, because the limit is structural — one reply, per comment, within seven days — rather than numerical.
That is the core reason official-API automation does not trigger the restrictions that password-based bots do. You are not sending unsolicited messages faster than a human could. You are sending exactly one reply to each person who asked you a question in public.
More on the mechanism in how the private reply API works, and on the seven-day rule specifically in the seven-day reply window.
The 24-hour window, which is not the same thing
Worth separating, because it catches people out.
The seven-day window governs private replies to comments. The 24-hour window governs ordinary conversation: once someone sends you a message, you can reply freely for 24 hours. After that, standard messaging closes and only specific message types are permitted.
So a person who comments, receives your private reply, and then writes back has opened a 24-hour conversation window — which is why the follow-up question matters so much in every campaign structure worth running. It is not just engagement; it is the mechanism that keeps the conversation open at all.
Rate limits, and when you will meet them
Meta enforces limits on how many API calls an app can make in a given period. In normal operation you will never approach them.
The exception is precisely the situation you want: a reel takes off and 3,000 comments land in ninety minutes.
What separates tools here is what they do at that moment. A well-built one queues and sends steadily inside the limits until the backlog clears — everyone gets their DM, some a few minutes later than others. A badly built one fires everything at once, gets throttled, and drops the remainder, so the people who commented during your single best hour receive nothing at all.
Ask about queue behaviour before you need it. It only becomes visible on the day it matters most.
The third kind of limit: your tool's plan
Entirely separate from anything Meta enforces, and the one most likely to actually stop your automation.
Tools cap usage by plan, and the model differs in ways that matter:
- Per active contact — you pay for unique people per month. ManyChat works this way, at 250 contacts on its entry tier with overage above that. A post doing unexpectedly well is therefore also your most expensive one. The maths is in ManyChat pricing explained.
- Per DM — a monthly message allowance. Predictable, but a viral post can exhaust it.
- Per connected account — flat, regardless of volume. Viral posts cost nothing extra.
When your automation stops, check this before assuming Instagram restricted you. In most cases it is a plan cap, not the platform.
If you have actually been restricted
Stop sending immediately. Continuing while restricted extends it.
Work out which behaviour caused it. Manual bulk outreach, a password-based tool, or buying engagement are the usual three. Official-API comment replies are rarely the cause.
Disconnect anything that logs in with your password. Change the password, then check authorised apps in your Instagram settings and remove what you do not recognise.
Wait it out. Most restrictions are temporary and lift on their own. Appeal through the in-app flow, once, and then leave it.
Do not immediately resume. Return to normal activity levels for a week or two first.
The short version
Nobody outside Meta knows the manual DM thresholds, and they are behavioural rather than numerical — identical messages to strangers is what causes trouble, not volume.
The API has no daily cap at all. It has structural limits instead: one reply per comment, seven days, professional account, no cold messaging.
And if your automation stops working, look at your tool's plan limits before you assume the platform did it. That is almost always the answer.