Skip to content
CommentGate
Log in Start free for 7 days

Comments that look after themselves.

Flip a few switches and write one line about your brand. It handles the rest, and every correction you make teaches it.

Start free for 7 days See how it works

Home

Handled 96% on its own this week +5 pts asked you about 86 agrees with your calls 98%

Handling scams and spam on its own now.

Needs you

3
  • Derek Olsen Facebook 4m

    Sole came apart on my second hike. Emailed support twice with no answer, disputing the charge if I do not hear back by Friday

    About chargebacks

  • Priya Raman YouTube 11m

    I review hiking gear for about 240k people on YouTube and would love to test the Ridgeline. Who do I talk to?

    About partnerships

  • @emma.on.trails Instagram 18m

    My podiatrist told me to try these and my plantar fasciitis pain is completely gone

    About health claims

A handful of switches. That is the setup.

Scams, spam, hate and competitor mentions are on from the start. Profanity, links, phone numbers, pictures and negative comments are one flip away.

  • One line of guidance Anything it should know about your brand, in plain words.
  • Stricter where it matters A different set of switches for one account when you need it.

Setup

Hide scams and fake giveaways Fake prize winners, "DM me to claim", fake support accounts and phishing.
Hide spam Unrelated promotion, follow-for-follow, adult spam and people plugging their own pages.
Hide hate and harassment Slurs, personal attacks on your customers or staff, and threats.
Hide competitor mentions People steering your customers to another brand.
Hide links Any comment with a link to another site. Links to your own site stay up.

It only asks about the hard ones.

Obvious spam is hidden and simple questions are answered without you. A chargeback threat, a partnership offer or a reporter waits in Needs you, with the reason it is asking.

  • Hide, Keep or Reply One click each, and the comment is done.
  • Everything else in Activity Every decision it made on its own, with the reason.

Activity

Every correction is a lesson.

Unhide a comment, keep one it would have hidden, or change a draft, and it remembers. When it agrees with your team 95 times out of 100 on a kind of comment, it stops asking about that kind.

  • You can see it learn Home shows which kinds it handles alone and which it still asks about.
  • It tightens up by itself If it starts getting a kind wrong, it asks again straight away.
  • Weekly suggestions Patterns in your corrections become one-click changes.

What it handles alone

Scams and fake giveaways Agrees with your team 99% of the time
On its own
Spam Agrees with your team 97% of the time
On its own
Hate and harassment Agrees with your team 96% of the time
On its own
Competitor mentions Agrees with your team 88% of the time
Still asking

Also part of the autopilot

  • Your own topics

    Describe anything else to catch in a sentence, like "questions about the recall".

  • Lists that always win

    Blocked and allowed words, trusted people and muted people.

  • Test first

    Paste any comment and see exactly what would happen, and why.

  • Answer questions

    Off, draft for you, or send clear answers on their own.

  • Prices from your catalog

    Price, stock and size answers come from your Meta catalog or product feed.

  • Last 30 days on connect

    Recent comments come in when you connect, so it starts with context.

  • Undo anything

    Hidden comments come back with one click on every platform.

Where it works

The same switches on every platform you connect.

  • Facebook Every switch, on posts and ads.
  • Instagram Every switch, on posts, Reels and ads.
  • TikTok Every switch, on videos and ads.
  • YouTube Every switch, on videos and Shorts.
  • Threads Every switch, on posts and Threads ads.

More on Facebook and Instagram, TikTok, YouTube and Threads.

Keep reading

Questions

How long before it runs on its own?

It works from the first comment with the switches you picked. It asks about fewer comments as it sees your corrections, and once it agrees with your team 95 times in 100 on a kind of comment (after at least 50 answers), it handles that kind alone.

What does it ask me about?

Only the comments a person should decide: a customer threatening a chargeback, a partnership offer, a reporter, a safety report. Spam, scams and simple questions are handled without you.

What if it hides something it should not have?

Unhide it with one click from Activity. The comment comes back, and it learns from the correction.

Can I still write my own rules?

Yes. Rules with explicit actions and your lists always win over the switches, so you can be as precise as you like on any post or ad.

Does it work in other languages?

Yes. It reads comments in any language and can draft replies in the language they were written in. It is most confident in English, so it asks about uncertain comments in other languages a little more often.