Ask Claude or ChatGPT about your comments
The Social Moderator AI MCP server connects your workspace to AI assistants that support the Model Context Protocol. Ask what needs your attention, pull up hot leads or read why a comment was hidden. When you ask, your assistant can also draft and post replies, hide spam and block trolls, all from the chat you already use.
Connect in two minutes
-
Copy the server URL
This is the address your assistant connects to.
https://socialmoderator.ai/mcp -
Add it to your assistant
Add it as a custom connector in Claude, or in ChatGPT with developer mode turned on. Claude Code, Cursor and VS Code can add it as a remote MCP server.
-
Sign in and approve
Your assistant opens the Social Moderator AI sign-in page. Sign in, approve access on the consent screen, and it works with the workspace you have selected in the app.
-
Ask
Start with “What needs my attention today?” and go from there.
What your assistant can see
- What needs attention
- Comments waiting for review, open hot leads, reply drafts, how many Google reviews and DMs need an answer, and comment counts per page for the last 7 days.
- Every comment, with the reasoning
- Search the inbox by status, page or text. For any comment, see the post it was left on, the suggested action, the rules it matched and any reply drafts.
- Leads, scored
- Comments from people asking about buying, booking or pricing, each with a score from 0 to 100, the buying intent and whether the lead is still hot.
- Your rules and brand voice
- Read the live moderation rules and brand voice, and check how the AI would handle a comment before one like it shows up. Nothing is posted.
What it can do when you ask
- Draft replies in your voice
- Write a reply in your brand voice and save it as a draft on the comment. Nothing is posted until you approve it.
- Post the replies you approve
- Post a reply to a comment, or apply the action the moderation AI suggested, once you confirm it in the chat.
- Hide and unhide comments
- Hide spam, scams and abuse on your pages, and bring back a comment that was hidden by mistake.
- Block a troll
- Block a comment’s author so their future comments are hidden without asking the AI, and unblock them when you change your mind.
By default, Claude and ChatGPT ask you to confirm before they post a reply, hide a comment, apply the AI’s suggestion or block someone.
Things to ask
- What needs my attention today?
- Show me the open leads from this week.
- Why was the latest comment on our Instagram hidden?
- Would our rules hide “DM me for a discount code”?
- Draft a reply to the wholesale question and post it after I approve.
- Hide the spam comment on our latest reel.
Common questions
What is MCP?
The Model Context Protocol is an open standard that lets AI assistants use tools from other apps. Claude, ChatGPT, Cursor and many other clients support it.
Can my assistant change my rules or settings?
Only the blocked accounts in your rules, when you ask it to block or unblock a commenter. The rule text, brand voice, connected pages and billing stay in Social Moderator AI: your assistant can read your rules and brand voice and test a comment against them, but not edit them.
Which workspace does my assistant see?
The one you have selected in Social Moderator AI. Switch workspaces in the app and the assistant follows.
How do I disconnect it?
Remove the connector in Claude, ChatGPT or your other MCP client. You can also revoke access in Social Moderator AI under Settings > Connected apps.
Are my comments used to train AI models?
Social Moderator AI never uses your comments to train AI models. What your assistant does with the answers it receives is governed by that assistant’s own privacy settings.
Where do I get help?
Email info@socialmoderator.ai or use the contact form at https://socialmoderator.ai/contact. We reply within one business day.