MCP Setup

Connect Zilfu to Claude, Cursor, ChatGPT, and other AI assistants over the Model Context Protocol.

What is MCP?

The Model Context Protocol (MCP) is an open standard that lets AI assistants like Claude, Cursor, and ChatGPT talk directly to your tools. Instead of clicking through dashboards, you describe what you want in plain language and your AI assistant takes care of it for you.

Zilfu ships an MCP server, so any MCP-compatible client can read your data, create content, and schedule posts on your behalf — using the same personal access token that powers the REST API.

What You Can Do

Through a normal conversation with your AI assistant, you can:

Find your workspaces and accounts — list every space you can publish to, and the social accounts connected to each.
Create and schedule posts across one or many accounts in a single request, including multi-account clusters.
Manage your weekly schedule — list, create, and delete recurring posting slots.
Inspect your account — confirm who you're authenticated as before taking action.

The MCP server currently exposes the following tools:

Tool	What it does
`ping`	Health check — confirms the connection is live
`get-account`	Returns the authenticated user account
`list-spaces`	Lists every workspace you have access to
`list-accounts`	Lists connected social accounts in a space
`list-slots`	Lists the recurring weekly posting slots in a space
`create-slots`	Creates one or more recurring slots
`delete-slot`	Deletes a recurring slot
`create-posts`	Creates posts (immediate, scheduled, or queued in a slot)

More tools (media uploads, post listing and editing, analytics) are on the roadmap — the REST API already covers them today, see the API Reference.

Before You Start

You need a personal access token. If you don't have one yet, follow the Quickstart — it walks through creating one from Settings → API Tokens.

The MCP server endpoint is:

https://zilfu.app/mcp/social-media

Authenticate every connection by sending your token as a Bearer token in the Authorization header. The same token works for both the REST API and the MCP server.

Treat your token like a password. Anyone with it can act on your account. Never commit it to source control or paste it into shared chats.

Connecting Your Client

The exact setup depends on the AI assistant you use. Pick the one that matches your workflow.

Claude Code

From your terminal:

claude mcp add --transport http zilfu \
  https://zilfu.app/mcp/social-media \
  --header "Authorization: Bearer YOUR_TOKEN"

Then start a session with claude and ask it to list your spaces — you should see the zilfu server listed in /mcp.

Claude Desktop

Open Settings → Developer → Edit Config and add the Zilfu server to your mcpServers map. Claude Desktop reaches remote HTTP servers through the mcp-remote proxy:

{
  "mcpServers": {
    "zilfu": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://zilfu.app/mcp/social-media",
        "--header",
        "Authorization: Bearer YOUR_TOKEN"
      ]
    }
  }
}

Restart Claude Desktop. The Zilfu tools will appear in the tools menu of any new conversation.

Cursor

Open Settings → MCP → Add new MCP Server and paste the same JSON snippet shown above for Claude Desktop. Cursor uses the same configuration shape and the same mcp-remote proxy.

Other MCP Clients

Any client that speaks MCP can connect. Point it at https://zilfu.app/mcp/social-media over streamable HTTP with an Authorization: Bearer YOUR_TOKEN header. Clients that only support the local stdio transport can reach the server through the mcp-remote proxy, exactly like the Claude Desktop example above.

Verifying the Connection

Once your client is connected, ask the assistant something concrete to confirm everything works end-to-end:

"List my Zilfu spaces."

The assistant should call list-spaces and respond with the workspaces attached to your account. From there, ask it to draft a post, fill out your weekly schedule, or queue content into the next available slot — all in plain language.

Troubleshooting

401 Unauthorized — your token is missing, malformed, or has been revoked. Generate a new one from Settings → API Tokens and update your client config.
Tool not visible — make sure the client has been restarted after editing its config, and confirm the server name (zilfu) doesn't collide with another entry.
Nothing happens when you ask — check that the assistant actually has permission to call MCP tools. Some clients require you to enable tool use per-conversation.

On this page