# Facebook MCP server: manage your Page from Claude or ChatGPT

> Yes: HeyMark's MCP server connects your Facebook Page to Claude, ChatGPT or any MCP client that supports OAuth. Your assistant schedules posts, Reels and Stories, reads the comments and messages in your inbox, replies after you approve and builds automations that send a DM to whoever writes a keyword.

By HeyMark · Reviewed by Crescente Martinez, CMO and co-founder of HeyMark · Updated September 25, 2026

You ask ChatGPT "who messaged the Facebook Page about reservations since Friday? Draft replies and wait for my confirmation". Your assistant goes through the comments and messages in the inbox, drafts the replies and sends them once you approve. That's HeyMark's Facebook MCP server: it lets you manage your Facebook Page from Claude, ChatGPT or any MCP client that supports OAuth, under your own login and permissions.

If people message your business on Facebook to ask about hours, prices or bookings, the way they do with a restaurant or a dental clinic, this page starts with the inbox; publishing comes after. MCP access is free on every plan.

## What your assistant can do on Facebook

### Comments and messages in one inbox

Your Page's comments and direct messages land in HeyMark's inbox, next to those from your other connected accounts. Your assistant can:

- see how many unread messages there are and which conversations are waiting on a reply;
- read the full history of a conversation;
- draft replies and send them once you approve;
- label conversations, mark them as read and pause AI in one of them;
- review the conversations escalated to a person and resolve them.

### Automations and AI replies

Both keyword rules work on Facebook: the comment rule, which replies publicly and sends a DM to whoever comments the word, and the DM rule, which answers whoever sends the word by direct message. Each rule takes up to 20 keywords and uses no credits. If you'd rather have AI write the replies instead of fixed texts, AI DM automation answers in your brand's voice and is set per brand; each AI reply uses 1 credit. How a comment-to-DM rule is built, with Instagram as the example, is covered in [Instagram DM automation](https://heymark.ai/en/solutions/instagram-dm-automation/).

### Publish and schedule

It creates posts, Reels and Stories, and schedules them or publishes them right away. Every piece sits on the HeyMark calendar with a status, a date and an owner.

### Metrics

Facebook is part of HeyMark's analytics overview: followers, views and interactions against the previous period, plus metrics for each post.

## What needs your approval

Replying to a comment or a message, publishing right away, creating or changing an automation and changing the AI DM settings show you a preview to approve first. Pausing AI in a conversation runs right away; resuming it asks for confirmation. Scheduling also runs right away, so check the date and copy before you ask. What gets published is your call.

## Prompts for Facebook

#### Messages waiting on a reply

```text
Which Facebook conversations in the HeyMark inbox are waiting on a reply?
Sum up each one in a line and flag the ones asking about bookings, hours or prices.
Draft a reply for each and don't send anything until I approve each message.
```

#### Comment-to-DM for bookings

```text
Create an automation for Facebook: if someone comments BOOK, reply publicly with "We sent you a message" and send them this link by DM: [booking link].
Show me the rule before turning it on.
```

#### Review the AI for DMs

```text
Read my brand's AI settings for direct messages.
Tell me what business information it's missing to answer questions about hours, prices and bookings well, and draft the text you'd add.
Don't change anything without my approval.
```

#### Conversations handed to a person

```text
Which Facebook conversations were escalated to a person this week?
Sum up each one and tell me which are still open.
```

#### The weekend post

```text
Create a post for the Facebook Page with the image I pick in HeyMark's file picker and this text: [text].
Show me the draft and wait for my confirmation before scheduling it for Saturday at 11 a.m.
```

## How to connect it

1. In HeyMark, open My accounts in your brand and connect Facebook.
2. Add the URL `https://mcp.heymark.ai` to your assistant: in Claude as a custom connector ([guide for Claude](https://heymark.ai/en/integrations/claude/)) and in ChatGPT with developer mode ([guide for ChatGPT](https://heymark.ai/en/integrations/chatgpt/)). The Claude Code and Codex commands are in the [MCP documentation](https://heymark.ai/en/mcp/).
3. Approve the access when HeyMark opens in your browser, then ask your assistant which Facebook account it sees and how many conversations are waiting on a reply.

## Limits

- It isn't an official Meta MCP server. It uses the account you connected in HeyMark, with your permissions.
- It doesn't manage paid ads: boosting a post stays outside HeyMark.
- Each brand connects one Facebook account. If you run Instagram too, its comments and messages land in the same inbox; what's specific to that network is in the [Instagram MCP server](https://heymark.ai/en/mcp/instagram/).
- Scheduling has no preview: a scheduled post goes out on time.
- Facebook accepting the request doesn't mean the post is live. Ask your assistant to check the actual status.
- There's no public REST API yet.

## Frequently asked questions

### Is this an official Meta MCP server?

No. It's HeyMark's MCP server. It uses the Facebook account you connected in HeyMark, with your permissions, and your assistant sees what you see and nothing else.

### Can it answer my Page's messages?

Yes. Your assistant reads the inbox conversations and sends the reply once you approve the preview. If you turn on AI DM automation, Mark answers in your brand's voice and escalates conversations that need a person.

### Does it handle Facebook ads?

No. HeyMark doesn't manage paid ads: it works with your posts, your inbox and your automations.

### How much does it cost?

MCP access is free on every HeyMark plan; Creator costs $12 USD a month, and every plan starts with a 14-day free trial. An AI reply or DM uses 1 credit, keyword automations use none, and your AI client may charge separately.

### Do I need to know how to code?

No. You paste the server URL into Claude or ChatGPT, sign in with your HeyMark account and approve the access. There are no keys to generate.

## See it on your own accounts.

Connect Instagram, TikTok, Facebook, LinkedIn or YouTube and use HeyMark free for 14 days. Mark answers in your voice, and you approve what goes out.

[Start free](https://app.heymark.ai) · [See pricing](https://heymark.ai/en/#pricing)

## Keep reading

- [Instagram MCP server: run your account from Claude or ChatGPT](https://heymark.ai/en/mcp/instagram/): An Instagram MCP server for Claude and ChatGPT: schedule Reels and carousels, reply to comments and DMs after you approve, and build comment-to-DM automations.
- [Instagram DM automation: auto DM everyone who comments your keyword](https://heymark.ai/en/solutions/instagram-dm-automation/): Instagram DM automation for comments: anyone who comments your keyword gets a public reply and a DM with your link. On every HeyMark plan, no credits spent.
- [Connect Instagram to Claude to post, reply and analyze](https://heymark.ai/en/integrations/claude/): Connect Instagram to Claude with HeyMark's custom connector: analyze your posts, plan the month, schedule, and reply to comments and DMs with your approval.
- [Connect Instagram to ChatGPT to post and reply](https://heymark.ai/en/integrations/chatgpt/): Connect Instagram to ChatGPT through HeyMark so it can build carousels, schedule, publish with your approval and answer comments and DMs from the chat.

## Developer and agent resources

- [HeyMark MCP documentation](https://heymark.ai/en/mcp/)
- [llms.txt](https://heymark.ai/llms.txt)
- [Full site content for language models](https://heymark.ai/llms-full.txt)
- MCP protocol endpoint: `POST https://mcp.heymark.ai`
- OAuth protected-resource metadata: [/.well-known/oauth-protected-resource](https://mcp.heymark.ai/.well-known/oauth-protected-resource)
- MCP server card: [/server-card](https://mcp.heymark.ai/server-card)
