Model Context Protocol · JSON-RPC 2.0

Turn your AI assistant into a social media manager

Connect Postger's MCP server to Claude, ChatGPT, Cursor, or any AI assistant and manage your whole social presence in plain English. Draft and schedule posts, search your media library, and read analytics across 9 platforms — without ever opening a dashboard.

Works withClaudeChatGPTCursorGeminiCopilotPerplexity

No credit card required · Included on every plan

Claude · Postger connected
Draft a launch post for our new analytics dashboard and schedule it to LinkedIn and X for Thursday at 9am.
list_accountscreate_postschedule_post

Done — I scheduled your launch post to LinkedIn and X for Thursday at 9:00 AM. Here's the draft:

LinkedInXQueued · Thu 9:00 AM

📊 Your numbers, finally in one place. Today we're launching the new analytics dashboard — track every post and channel at a glance. See what's working, do more of it.

The real experience — your assistant calls Postger's MCP tools and schedules the post for you.

.cursor/mcp.json
{
  "mcpServers": {
    "postger": {
      "url": "https://app.postger.com/api/v1/mcp",
      "headers": {
        "Authorization": "Bearer pg_live_..."
      }
    }
  }
}

Using Claude Desktop? Just add the URL as a custom connector — no API key needed, OAuth is automatic.

How it works

1

Ask in plain English

Type what you want — “Draft a launch post for Thursday and queue it to LinkedIn and X.” No dashboard, no clicking around. Your assistant understands the request and knows which Postger tools to use.

2

Your AI runs the tools

Postger's MCP server exposes a focused set of tools. Your assistant lists your connected accounts, drafts the post, attaches media from your library, and schedules it — all using the scoped permissions you granted, nothing more.

3

It's done

The post lands in your Postger queue and publishes at the scheduled time through each platform's official API. You get a confirmation right in the chat, and every action is written to an audit log you can review.

Built for AI agents

The MCP server is designed for assistants to plan and act reliably — structured, permissioned, and model-agnostic.

Structured output

Every tool returns clean JSON, so assistants can parse results, plan next steps, and chain actions.

Model-agnostic

Works with any MCP client — Claude, ChatGPT, Cursor, Claude Code — no SDK or lock-in.

Scoped & auditable

Grant draft-only or publish access, restrict to accounts, and review every action in an audit log.

Human in the loop

Keep a person in control: use draft-only keys so nothing goes live without your explicit approval.

One connection, 9 platforms

Every platform Postger supports is available through the MCP server, each via its official API. Fan a single instruction out across your whole presence.

Instagram
X (Twitter)
Facebook
LinkedIn
YouTube
TikTok
Threads
Bluesky
Pinterest

The tools your assistant gets

Postger's MCP server exposes a focused toolset covering the full posting lifecycle. Each tool requires a specific permission — your assistant only sees the tools your key allows.

ToolWhat it does
list_accountsList every connected social account and its channel ID.
create_postCreate a draft or queued post with content, media, and a schedule.
list_postsList posts, optionally filtered by status (draft, queue, published).
get_postFetch a single post by ID.
update_postEdit an existing post's content, schedule, or images.
schedule_postQueue a post for publishing at a chosen time.
cancel_postPull a scheduled post back to draft.
upload_mediaUpload an image or video into your media library.
search_mediaSearch your media library by file name and type.
get_mediaFetch a single media asset by ID.
get_account_analyticsRead account metrics over a date range.
get_post_analyticsRead engagement metrics for a specific post.

What is an MCP server for social media?

MCP stands for Model Context Protocol — an open standard that lets AI assistants like Claude and ChatGPT securely connect to external tools and act on your behalf. Instead of copying and pasting between a chat window and your dashboard, the assistant talks directly to the service through a small, well-defined set of tools. Think of it as a USB-C port for AI: one standard connection that any compatible assistant can plug into.

Postger's MCP server turns your AI assistant into a social media manager. Once connected, you can ask your assistant to draft captions, schedule posts, reschedule your week, search your media library, and pull analytics — in plain English, without ever opening the Postger dashboard. The assistant does the clicking; you just describe what you want.

This matters because the fastest-growing way people work is now inside a chat window. If your team already lives in Claude, ChatGPT, or Cursor, the MCP server meets them there. There is nothing new to learn and no context-switching. Your social workflow becomes a conversation.

How Postger's MCP server works

The MCP server runs at a single endpoint and speaks JSON-RPC 2.0, the transport that MCP clients expect. When your assistant connects, it asks the server which tools are available, and Postger returns only the tools your credentials allow. From there, every request the assistant makes — creating a post, reading analytics — is a tool call routed through your workspace with your permissions applied.

Nothing about the underlying automation changes. A post created over MCP goes into the exact same queue as a post created in the dashboard, publishes through each platform's official API at the scheduled time, and appears on your calendar alongside everything else. The MCP server is simply a second front door into the product — one built for AI assistants instead of humans.

Because MCP is an open standard, the same connection works across assistants. Set it up once and use it from Claude on your laptop, ChatGPT on the web, and Cursor in your editor. You are not locked into a single vendor's ecosystem.

What your AI assistant can do

The server exposes a focused toolset covering the full posting lifecycle. Your assistant can list your connected accounts and their IDs, create drafts and queued posts, list and fetch existing posts, update content or scheduling, queue a post for publishing, and cancel a scheduled post back to draft. That is everything needed to plan and manage a content calendar from chat.

Media is covered too. Your assistant can upload images and videos into your media library and search that library by file name and type, so it can reuse an asset you already have instead of asking you to re-upload it. When you say “attach the product shot from last week,” it can find it.

For reporting, the server exposes account analytics over a date range and per-post engagement metrics. Ask “how did last month's LinkedIn posts perform?” and your assistant reads the numbers directly, so you can turn raw metrics into a summary or a client report without leaving the conversation.

Publish to 9 platforms from one conversation

Postger connects to Instagram, X (Twitter), Facebook, LinkedIn, YouTube, TikTok, Threads, Bluesky, and Pinterest — all through each platform's official API. Every one of those connections is available through the MCP server, so a single instruction to your assistant can fan a post out across your whole presence.

Ask for “a short announcement on X and Threads, and a longer version on LinkedIn,” and your assistant creates the right posts for each network in one go. Because Postger handles per-platform formatting and publishing under the hood, you describe the intent once and let the tools handle the platform-specific details.

This is the payoff of a first-party integration rather than an aggregator: your assistant works with real, native connections. Posts appear exactly as if you published them from each app, with full support for images, video, and platform-specific behavior.

Security: scoped, revocable, and auditable

The MCP server never sees your social passwords. Access is granted through a scoped API key or an OAuth token, and both carry only the permissions you choose — read posts, write posts, publish, read analytics, and so on. Effective permissions are the intersection of what the key grants and your current workspace role, so demoting a teammate automatically shrinks the reach of any key they issued.

You decide exactly how much rope your assistant gets. Create a key that can draft but not publish, and the assistant can prepare your week without anything going live until you approve it. Restrict a key to a single client's accounts, and it physically cannot touch the others. Access is revocable at any time — pull the key or revoke the token and the connection dies instantly.

Every action an assistant takes is written to an audit log with a timestamp, the tool that was called, and the originating request. Combined with per-key and per-workspace rate limits, that gives you a complete, reviewable record of what your AI did on your behalf. Trust, but verify — and here you actually can.

Never sees your social passwords
Scoped API keys and OAuth tokens
Draft-only keys — no publish access
Restrict a key to specific accounts
Revoke access instantly, anytime
Full audit log of every action

Connect your AI assistant in three steps

Setup takes a couple of minutes. First, open Settings → API Keys in Postger (owner or manager role required), create a key, choose the accounts and permissions it may use, and copy the pg_live_ token — it is shown only once. Then add the MCP server to your assistant. Then start asking.

For Claude Desktop and claude.ai, you do not even need an API key: add https://app.postger.com/api/v1/mcp as a custom connector and Claude discovers the OAuth flow automatically, sending you to Postger's consent screen to sign in, pick a workspace, and approve the scopes. Access refreshes on its own and you can revoke it anytime.

For Cursor, Claude Code, and other MCP clients, point the client at the same endpoint with your API key as a bearer token. The configuration is a few lines of JSON. Full copy-paste snippets for every client live in the API & MCP documentation, linked below.

Who it's for

Agencies use the MCP server to run more accounts with the same team. Instead of clicking through dashboards for every client, an account manager can ask their assistant to draft a week of posts, schedule them to the right workspace, and pull last month's numbers for a report — all in chat, with a per-client key that cannot touch other accounts.

In-house teams and multi-location brands use it to keep publishing consistent without adding headcount. Marketers who already draft copy in Claude or ChatGPT can push the finished post straight into the Postger queue, closing the loop between writing and scheduling in the same window.

Creators and founders use it to move fast. When your whole workflow is a conversation, maintaining a posting rhythm stops feeling like a chore. You think of an idea, describe it, and it is scheduled — no tab-switching, no forgetting, no friction.

Explore more Postger features

Frequently asked questions

What is an MCP server?

MCP (Model Context Protocol) is an open standard that lets AI assistants connect to external tools and act on your behalf. An MCP server exposes a set of tools that assistants like Claude and ChatGPT can call. Postger's MCP server exposes tools for drafting, scheduling, and managing social media posts, so you can run your social workflow from a chat window.

Which AI assistants work with Postger's MCP server?

Any MCP-compatible client works, including Claude (Desktop and claude.ai), ChatGPT, Cursor, Claude Code, and other assistants that support the Model Context Protocol. Because MCP is an open standard, you set the connection up once and use it from any compatible assistant.

Which social platforms can I manage over MCP?

The MCP server can manage all of Postger's connected platforms: Instagram, X (Twitter), Facebook, LinkedIn, YouTube, TikTok, Threads, Bluesky, and Pinterest. Every connection uses each platform's official API.

What can the AI actually do?

It can list your connected accounts, create draft and queued posts, list and fetch posts, update content and scheduling, queue posts for publishing, cancel scheduled posts, upload and search media, and read account and post analytics — all within the permissions you grant.

Is it safe to give an AI assistant access to my social accounts?

Yes. The MCP server never sees your social passwords. Access is granted through a scoped API key or OAuth token that carries only the permissions you choose, and effective permissions are limited by your workspace role. You can create a key that can draft but not publish, restrict a key to specific accounts, revoke access at any time, and review every action in an audit log.

Do I need an API key to connect Claude?

No. For Claude Desktop and claude.ai, add https://app.postger.com/api/v1/mcp as a custom connector and Claude discovers the OAuth flow automatically. You sign in, pick a workspace, and approve the scopes on Postger's consent screen. For Cursor, Claude Code, and other clients, you use an API key as a bearer token.

How do I connect Cursor or another MCP client?

Add the endpoint https://app.postger.com/api/v1/mcp to your client's MCP config with your API key in the Authorization header. For Cursor, that is a short block in .cursor/mcp.json. Full copy-paste configuration for every supported client is in the API & MCP documentation.

Can I limit what the assistant is allowed to do?

Yes, and it is recommended. Create a dedicated key per assistant with only the permissions it needs — for example, drafting without publish access — and restrict it to specific accounts. The assistant physically cannot perform actions outside those grants, and you can watch what it does in Settings → API Keys → Recent activity.

Are posts created over MCP the same as posts created in the dashboard?

Yes. A post created through the MCP server goes into the same queue, appears on the same calendar, and publishes through the same official platform APIs as a post created by hand. The MCP server is just a second way in — built for AI assistants.

Is the MCP server included in my plan?

Yes. The MCP server and REST API are available on all plans at no extra cost. Postger offers a 7-day free trial with full access; paid plans start at $17/month billed yearly, and the Agency plan at $41/month includes unlimited users, workspaces, and social profiles.

Are there rate limits?

Yes. Requests are limited per key and per workspace to keep the service stable — for example, write and read limits per minute per key, and a higher write limit per workspace. Throttled responses return a 429 status with a Retry-After header so your client can back off gracefully.

What is the difference between the MCP server and the REST API?

They expose the same capabilities through different interfaces. The REST API is for code and custom integrations; the MCP server is for AI assistants, which discover the available tools automatically and call them in response to natural-language requests. Both live under /api/v1 and use the same scoped API keys and permissions.

Ready to run social from your AI assistant?

Connect Postger to Claude, ChatGPT, or Cursor in two minutes and start scheduling in plain English.

$0.00 due today, cancel anytime