Instagram Comments Held for Review: Release Hidden Replies
TL;DR
Instagram's Hidden Words setting routes replies with flagged terms, or terms Meta's classifier marks as offensive, to a hidden folder on each post. The comments are not deleted. Open the post, tap View hidden comments, and release or delete each one. If your reach also dropped or the account is under review, the filter is not the whole story.
Why Instagram holds comments for review
If a comment never shows up in your notifications, or a follower tells you their reply disappeared from your post, Instagram did not necessarily delete it. In most cases the reply is still there, routed to a hidden area on the post itself, waiting for you to look.
The filter responsible is called Hidden Words. According to Meta's own announcement, it automatically filters comments that may be inappropriate or offensive, and it also catches terms you add yourself. Replies caught by either path do not reach the main comments list. They sit in a separate view that only the account owner can open.
This is a feature, not a glitch. The point is to protect the person running the account from targeted abuse without silently losing legitimate replies. The downside, which brings most people to this article, is that normal comments get caught too, especially around emoji chains, product mentions, and anything that resembles a scam pattern.
Where held comments actually live
Hidden comments do not appear in your notifications tab, in the comment thread under the post, or in the activity view. They are kept per post, inside a view labeled View hidden comments at the end of the comment list on that specific post.
If a reply has been held for review, three things are true at once:
- The commenter sees their comment as posted and visible to them
- Nobody else sees the comment unless you release it
- No notification is sent to you about it
That combination explains why creators and businesses only find out when a follower asks why their comment vanished. The platform does not surface the hidden queue proactively.
How to release a held comment
The steps work the same on iOS and Android. The web interface has a reduced comment view and does not currently expose the hidden list, so use the app.
- Open the post where the comment was expected
- Tap the comment icon to open the full comment list
- Scroll to the bottom of all visible comments
- Tap View hidden comments
- Review each entry, then tap the three dots next to a specific comment
- Choose Unhide to release it, or leave it hidden to keep it filtered
Unhiding does not notify the person who commented. From their side nothing changes, their comment stayed visible to them the whole time. For your audience, the comment then appears in the thread as normal.
The two controls that decide what gets held
Hidden Words has two separate switches that stack on top of each other. If you only turn one off, the other may still catch replies.
Offensive content filter. This is Meta's own classifier, tuned to patterns the platform considers inappropriate. You cannot see the list of terms, and the filter is on by default on many accounts. It also catches intentional misspellings and emoji substitutions, so a reply that looks fine to you may still contain a flagged pattern.
Custom word list. This is a list you build yourself, under Settings and privacy, then Hidden Words, then Manage custom words and phrases. Anything you add is filtered from comments and from message requests. The list is yours to clear or edit at any time.
If you cannot find a reason the reply was caught, open the custom list first. A word you added months ago, as part of a campaign or a mute of a running joke, may still be live and catching legitimate traffic.
Why legitimate replies get caught
Patterns that cross with Hidden Words detection, on accounts with meaningful follower volume, usually come down to one of these:
- Links or shortened URLs, which Meta tends to associate with promotional abuse
- Repeated emojis, especially flame, fire, or money related ones, which match attention-farming patterns
- Capitalised shouting phrases, which overlap with aggressive-reply patterns
- Keywords in common misspellings that match the classifier's offensive list, even when the writer meant something benign
- Comments posted immediately after other similar comments by accounts the platform has already restricted
None of this means the commenter did anything wrong. It means the filter is tuned toward false positives rather than false negatives, a design choice that favours safety at the cost of noise. Releasing legitimate comments individually is the current workaround.
When held comments are not the real issue
Comments being held is routine. Comments being held alongside other symptoms is a different problem, and you should treat it that way.
Signals that the account itself is in a restricted state, not just the comment filter:
- Posts stopped appearing on hashtag pages or on Explore
- Reach dropped by half or more across posts, not only on a specific one
- A banner at the top of the app says the account is under review, limited, or at risk
- You received a warning for a specific post, Reel, or Story
- You cannot comment, like, or follow from the account at usual rates
Each of these points at a different mechanism. A reach drop with no warning often points to a suppressed account rather than a comment filter. A warning on content you did not expect usually points to a strike against the account. A limited banner points at a direct restriction, and the standard in-app appeal is the first step before anything else.
Prevention and settings tuning
If you want fewer legitimate comments held in the first place, Hidden Words can be tuned. The feature lives under Settings and privacy, then Hidden Words.
- Toggle off Hide comments to disable the offensive content filter entirely. Replies still subject to platform-wide moderation will still be removed by Meta, but the per-account filter stops holding comments on your side
- Review the custom list and remove anything stale
- Keep Hide message requests on if you receive scam DMs, since this part of the setting is unrelated to comments on your posts
- Pin a comment that answers the most common question on your post, which reduces the volume of replies the filter has to score
None of this unhides a comment that is already held. Settings only change behaviour from the moment forward.
When the restriction crosses into a right of access
If the account itself is restricted, not only the comment filter, EU users have escalation routes beyond the in-app appeal. These apply to platform decisions that limit visibility, suspend features, or disable the account.
Under the Digital Services Act (Article 20), platforms must offer an internal complaint-handling system for at least six months after a restriction decision. Under Article 21, if the internal appeal does not resolve the dispute, you can bring it to a certified out-of-court dispute settlement body in your Member State. Under the GDPR (Article 15), you have the right to access information the platform holds about you and the processing applied to the account.
These are the three channels the law actually provides. They are not shortcuts to a faster review, they are formal routes with their own procedures. When the problem is a stuck comment filter and nothing more, none of them apply. When the account itself is limited, all three can come into play depending on what you want: a reversal, a review, or an explanation.
When professional recovery makes sense
For a filter issue, no outside help is needed. For a restriction, suppression, or disabled account, in-app appeals are the first step and often enough. When they are not, and especially when they return no case reference and no stated timeline, legal escalation is the route that remains.
Recover's legal team handles exactly this: individual case review for restricted and disabled Instagram accounts, built on GDPR and DSA arguments rather than resubmission of the same form. Current performance is a 97% success rate across cases, with 96% resolved within 30 days. A full money-back guarantee applies when recovery fails. See service tiers for the current terms.
If the account is healthy and only comments are being filtered, keep the fix to the Hidden Words setting. If multiple symptoms stack, the restriction, not the filter, is what to act on.