Before you pay for any comment-moderation tool, you should know what Meta already gives you for free — and use all of it. This guide explains Facebook's built-in moderation settings honestly: what each one does, where to find it, and the specific places where keyword-only filtering stops being enough for paid ads.
One caveat up front: Meta reorganises these settings regularly, and what you see depends on whether your Page is in the classic layout, the new Pages experience, or professional mode. As of mid-2026 you will find most of them under your Page or profile Settings, in sections named along the lines of Comments, Hidden Words, or Moderation — often surfaced through Meta Business Suite or the professional dashboard. If a menu below has moved, search the settings for the feature name.
The profanity filter
The longest-standing tool. When enabled, Facebook automatically hides comments containing words from Meta's own profanity list — a list Meta maintains and does not publish, historically informed by commonly reported words and phrases. Older Page settings offered a strength setting; newer surfaces tend to present it as a simpler toggle folded into the Hidden Words controls.
What it is good at: keeping the crudest language out of public view with zero effort. What it cannot do: anything else. It has no opinion on scam links, crypto spam, competitor bashing, or "this brand is a ripoff" — none of which need a swear word.
Hidden Words and custom keyword lists
Hidden Words is the more capable successor, spanning Facebook and Instagram. It has two parts:
- Offensive content filtering. Meta auto-hides comments (and, depending on settings, message requests) that its systems consider offensive.
- Custom words and phrases. You supply your own list — words, phrases, emoji — and comments containing them are hidden automatically.
Hidden comments land in a hidden/filtered queue you can review, and you can unhide anything caught by mistake. A well-tended custom list is genuinely useful: seed it with your recurring spam patterns ("dm me", "check my page", currency symbols plus links, the usual crypto vocabulary) and it will quietly absorb a lot of junk.
The limits are structural. Keyword lists match strings, not meaning: "scam" catches "this is a scam" but also "is this a scam? asking because I want to order" — a buyer-intent comment you just buried. Spammers rotate spellings faster than you rotate lists. And the list is one-size-fits-all: the same words apply to every post and every ad, with no way to be stricter on a high-spend campaign than on your organic memes.
Moderation Assist
Facebook's Moderation Assist (available in professional mode and on newer Page surfaces) moves beyond keywords to criteria about the commenter: you can auto-hide comments from accounts with no profile photo, very new accounts, accounts with no friends or followers, comments containing links, and similar. Some of these criteria pair well with ads — throwaway spam accounts usually look exactly like what these rules describe.
It is worth enabling. But it, too, judges the account and the surface features of the comment, not what the comment means — and it offers no path from "comment kept" to "comment answered."
Where the built-ins stop for advertisers
Put the three tools together and you have a decent free floor. Here is what is still missing for paid campaigns specifically:
- No ad context. Meta's filters do not know a comment sits under a live ad, let alone which campaign, ad set, or creative — the same blunt rules apply to a dark post burning budget and a three-year-old organic photo. There is no per-campaign strictness, and no view of moderation by ad.
- Keyword matching, not classification. The filters cannot tell spam from sarcasm from a shipping question that happens to share a word with your blocklist. Intent is invisible to them.
- Nothing happens to the comments that matter. The whole apparatus is subtractive. The comment that says "how do I order?" sails through untouched — which is correct — and then nobody is told it exists. On ads, those unanswered questions are the expensive part; see how to turn ad comments into customers.
- Visibility gaps on dark posts. Filtered-comment review queues are built around your Page's surfaces. Comments on ad-only posts are easy to lose track of entirely — a problem in its own right, covered in why comments aren't showing on your Facebook ads.
- No spam-storm response. When a spam wave hits one ad, keyword lists do not tighten, nobody is alerted, and the backlog waits for you.
The layered setup that works
Treat moderation as layers rather than a choice:
- Turn on everything free. Profanity filter, Hidden Words with a maintained custom list, Moderation Assist criteria. This is the floor and it costs nothing.
- Add an ad-aware layer on top. ROAS Shield connects through Meta's official API and moderates the comments Meta's filters cannot judge: it classifies every new ad comment by meaning (spam, abuse, off-topic, neutral, buyer-intent), applies your per-workspace rules, hides junk within seconds — hide, not delete, by default — and routes buyer-intent comments to you with a drafted reply. AI replies are draft-only unless you explicitly enable auto-send, and it works across dark posts and every ad instance, because it maps comments to ads rather than to your Page timeline.
The built-ins subtract the obvious junk; the ad-aware layer protects spend and converts the questions. Neither replaces the other.
Bottom line
Use Meta's free tools — all of them, today. Then be honest about what they are: keyword and heuristic filters shared across your whole presence, blind to campaigns and to meaning. If comments on your paid traffic are a real revenue surface for you, that is the gap ad-level comment moderation exists to close — plans and limits are on the pricing page.