You can automate Instagram comment hiding with a workflow that runs before a comment reaches your public post. Creator OS matches keywords and can hide any comment through the API, so a bot does not keep a public slot on your post, and you or your agent make the hide decision explicitly.
new comment arrives
|
v
+---------------------------+
| read the comments |
| list_comments / inbox |
+---------------------------+
|
v
+---------------------------+
| keyword match? |
| exact / contains / word |
+---------------------------+
| |
v v
no match match
| |
v v
hide_comment reply_to_comment
(your call) + optional DM
| |
+----------+----------+
v
your public post stays cleanDoes Instagram hide comments automatically?
Instagram has automatic comment filtering that evaluates the text of a comment, and Creator OS adds a second layer you control.
The app’s hidden words list covers a lot of noise. Instagram offers default filters for common offensive terms, and you can add your own words and phrases. Comments that match are hidden from the post but still visible to the author, and you can review them under Hidden comments.
A generic bot comment like “nice post, check my page” carries no banned word, so word filters may not catch it. The signal that gives it away is often the account behind it, not the words. That is where a second layer helps: read the comments through the API, look at who is commenting, and decide per comment whether to hide it or answer it.
Creator OS gives you that second layer. It reads comments, matches keywords, replies and can hide any comment you point it at. The judgement about which authors look empty stays with you or with your agent, which is safer than a rule you cannot see.
What Instagram comment spam from bots looks like in 2026
Bot comments have gotten quieter, which is why a workflow beats a single word list.
The obvious spam from a few years ago (crypto giveaways, “dm me for growth”) still exists, and Instagram’s default filters catch a lot of it. The softer versions you will meet in review:
- A single emoji on your Reel, from an account with no photo, no followers, and a link in bio.
- Generic praise (“Great content, this helped me so much”) on a post the account seems to have no history with.
- A friendly question that turns into a sales pitch in the DM.
- Several comments that all push another handle instead of talking about your post.
None of these are profane. All of them are cheap to spot once you look at the commenter as well as the comment.
Instagram spam comments vs real fans: the signals that separate them
You do not need a classifier. A few account level signals separate a bot from a fan, and you can read them from the comment list before deciding what to do.
| Signal | Bot pattern | Real fan pattern |
|---|---|---|
| Followers on the commenter | 0 | Any number, usually stable over months |
| Following (friends) count | 0 or thousands with no posts | Mixed, active |
| Profile photo | Missing | Present |
| Comment history across your posts | One comment, one post, gone | Repeats across weeks |
| Comment text | Templated, no reference to your content | References the post, asks a specific question |
The first two rows are the ones you act on. You pull the comments on a post, look at the authors, and hide the ones that look empty. There is a walkthrough of that pattern on the blog: hide comments if the author has no friends or followers.
How to automate Instagram comment hiding
Comment hiding in Creator OS is a called action, so build the read, match and hide steps into one loop that you or an agent runs.
Start with the reply side, which is fully automatic. Here is an automation created from the CLI:
creatoros automations:create \
--name "workshop-seat" \
--accountId acc_instagram_main \
--postId post_123 \
--keywords "seat,link" \
--matchMode word \
--commentReply "Just sent it to your DMs." \
--dmMessage "Here is the workshop link."That rule answers real questions and sends the link privately. Now the hiding step, which you or your agent run on any comment id:
creatoros inbox:hide-comment post_123 cmt_456 \
--accountId acc_instagram_mainAdd --unhide to reverse it if you hid a real fan by mistake. Nothing is deleted, so a false positive costs you one command, not a lost customer.
To see the raw material first, list what is waiting:
creatoros inbox:post-comments post_123 --accountId acc_instagram_mainFrom an AI agent, the same steps are tool calls: list_comments to pull the comments on a post, then hide_comment for the ones whose author looks empty, and reply_to_comment for the ones that need an answer. If you want the agent to explain its decisions in plain English before acting, the Social Agents docs cover the harness that runs your socials on Creator OS.
Rule 1: hide comments from accounts with no followers or friends
This is the single highest leverage review for a Creator or brand account, because spam accounts are structurally empty.
A working setup:
- Connect the Instagram account in Creator OS (Business or Creator account required).
- Pull the comments on the post with
inbox:post-commentsor your agent’slist_commentscall. - Look at the author of each comment. Empty profile, no followers, no following, no photo: hide it.
- Run
inbox:hide-commentor thehide_commenttool on each one. - Leave your Instagram default word filters on for profanity. The two layers do different jobs.
The result: a bot comments, you or your agent spot the empty account, and you hide the comment from public view with one command. Your notifications stay clean, and your comment section reads like a community instead of a classifieds page. The broader version of this workflow, including keyword funnels, is in hide spam comments automatically on Instagram.
Rule 2: pair hiding with Instagram auto replies so real comments get answers
Hiding without replying is half a system. The point is to remove noise and answer everything else, fast.
Creator OS supports comment replies on Instagram, Facebook, X, Threads, YouTube, LinkedIn and Reddit, and private replies on Instagram and Facebook. So the pattern is:
- Hide the comments whose authors look empty.
- Reply publicly to comments that match a keyword, like “price” or “link”.
- Send a private reply (a DM) to trigger a comment-to-DM funnel for a real buyer intent word.
For example, a creator selling a workshop sets the keyword “seat” to send a DM with the signup link, and publicly replies “just sent it to your DMs” so other readers see the answer exists. Bots never trigger it because they never use that word, and the empty accounts get hidden in review. The API side of comment to DM funnels is documented at comments to DM, and the AI flavored version is in reply to comments with AI.
Instagram auto reply vs auto responder vs AI comments: which do you need
These three terms get mixed up constantly, and picking the wrong one is why some automations feel robotic.
| Term | What it does | Best for |
|---|---|---|
| Auto reply | Posts a fixed public comment reply when a keyword matches | FAQ answers, “link in bio” requests, launch days |
| Auto responder | Sends a private reply or DM to a commenter | Lead capture, comment-to-DM funnels, support triage |
| AI comments | An agent reads the comment and writes a reply in your voice | Nuanced questions, multi language, high volume |
Most accounts need all three at different moments. A fixed auto reply is fine for “what is the price”. An auto responder handles the person who wants the link privately. An AI agent handles “does this work on iPhone 12 with a cracked screen”, where a template would look silly. Creator OS gives you keyword and DM automations plus MCP tools so a model can read a comment and reply, and the voice matching approach is covered in AI agent replies in your voice.
If you want to watch these flows built end to end, Kevin walks through them on the KevBuildsApps YouTube channel, and the overview video is here: Creator OS launch video.
FAQ
Does Instagram hide comments automatically?
Instagram auto-hides comments that match its default offensive filters or words you add to your hidden words list. That filtering works on the text of a comment, so it does not evaluate the account behind a generic “nice post” style comment. Creator OS adds a second layer where you read the commenter and hide what looks empty.
How do I stop spam comments on Instagram automatically?
Run two layers: keep Instagram’s hidden words on for profanity, then read comments through the API and hide the ones whose authors look empty. In Creator OS you list the comments, check the authors, and run the hide action for each one, either yourself or from an agent.
What is the difference between an Instagram auto reply and an auto responder?
An auto reply is a public comment answer posted under the post, for example a quick “sent you a DM”. An auto responder sends the private reply or DM, which is where links and offers belong. Creator OS can do both from the same keyword rule.
Can I hide comments from accounts with no followers?
Yes. Pull the comments with creatoros inbox:post-comments, look at who wrote each one, and run inbox:hide-comment on the empty accounts. You can undo it with --unhide, and an agent can do the same with the list_comments and hide_comment tools.
Do auto replies on Instagram hurt engagement?
Whether a reply helps depends on whether it fits the comment. A reply that names the thing the person asked about reads as a real answer, while a blanket reply that ignores the question reads as automated. Match the reply to the keyword and keep it human sized.
Get started
Create an account, connect Instagram, and build your first keyword rule and hide a spam comment in the same sitting. Creator OS plans start at $14.99/month for 11 platforms plus Skool, with the API, the hosted MCP server and the CLI included, so your agent can run the same steps without extra setup. Sign up at https://www.creatoros.ca/sign-up. If you want the API details first, read the docs at https://www.creatoros.ca/docs.
Get started for $0
Free trial for new subscribers. Cancel anytime.
