> ## Documentation Index
> Fetch the complete documentation index at: https://docs.postiz.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction

> Connect AI agents to Postiz using the Model Context Protocol (MCP)

MCP (Model Context Protocol) lets AI agents interact with Postiz directly, listing integrations, scheduling posts, generating images and videos, all through a standardized tool-calling interface.

This means you can connect Claude, ChatGPT, Cursor, or any MCP-compatible client to your Postiz account and manage your social media through natural language.

## How It Works

Postiz exposes an MCP server that provides **11 tools** to AI agents. The agent discovers these tools, understands their schemas, and calls them on your behalf.

```mermaid theme={null}
sequenceDiagram
    participant Agent as AI Agent
    participant MCP as Postiz MCP Server
    participant Postiz as Postiz Backend

    Agent->>MCP: Connect with API key / OAuth token
    MCP-->>Agent: List available tools
    Agent->>MCP: Call tool (e.g., schedule post)
    MCP->>Postiz: Execute action
    Postiz-->>MCP: Return result
    MCP-->>Agent: Tool response
```

## Available Tools

| Tool                   | Description                                                                             |
| ---------------------- | --------------------------------------------------------------------------------------- |
| `integrationList`      | List all connected social media accounts (optionally filtered by group)                 |
| `groupList`            | List all groups (customers) for your organization                                       |
| `integrationSchema`    | Get platform-specific posting rules and settings schema                                 |
| `triggerTool`          | Execute platform-specific helpers (e.g., list Discord channels)                         |
| `schedulePostTool`     | Schedule, draft, or immediately publish posts                                           |
| `postsListTool`        | List the organization's posts scheduled between two dates (with their current settings) |
| `postSettingsTool`     | Update the provider settings of a scheduled post or draft that was not published yet    |
| `generateImageTool`    | Generate AI images for posts                                                            |
| `generateVideoOptions` | List available video generation options                                                 |
| `videoFunctionTool`    | Get video generator settings (e.g., available voices)                                   |
| `generateVideoTool`    | Generate videos for posts                                                               |

## Authentication

There are two ways to authenticate with the MCP server:

### API Key

<Snippet file="api-key-where.mdx" />

Use it directly in the MCP endpoint URL, or as a Bearer token.

### OAuth Token

If you're building an app for other Postiz users, use [OAuth2](/public-api/oauth) to obtain tokens. OAuth tokens start with `pos_` and work the same way as API keys.

## Connecting

<Tabs>
  <Tab title="Bearer Token">
    Use the `/mcp` endpoint with your API key or OAuth token as a Bearer token:

    ```text theme={null}
    URL: https://mcp.postiz.com/mcp
    Authorization: Bearer your-api-key
    ```

    This method supports both API keys and OAuth tokens (prefixed with `pos_`).
  </Tab>

  <Tab title="API Key in URL">
    Use the `/mcp/:apiKey` endpoint with your API key embedded in the URL:

    ```text theme={null}
    URL: https://mcp.postiz.com/mcp/your-api-key
    ```
  </Tab>
</Tabs>

<Snippet file="base-url.mdx" />

## Quick Example

Here's what a typical interaction looks like when an AI agent uses Postiz MCP:

1. **Agent calls `integrationList`**, gets back your connected accounts (X, LinkedIn, etc.)
2. **Agent calls `integrationSchema`** with `platform: "x"` and learns X's character limits, settings, and rules
3. **Agent calls `schedulePostTool`**, schedules your post with the correct format

All of this happens automatically when you tell your AI agent something like:

> "Schedule a post to X for tomorrow at 10am: Excited to announce our new feature!"

## FAQ

### Do I need an OpenAI key to use Postiz MCP?

No. The MCP server just exposes Postiz's tools. Your AI client (Claude, ChatGPT, Cursor, etc.) provides the model. Postiz only needs an `OPENAI_API_KEY` if you use Postiz's own AI features (image generation, copilot) which are separate from the MCP tools surfaced to your client.

### What happens when my API key expires or is rotated?

Postiz API keys don't auto-rotate, but if you regenerate one in Settings → Developers → Public API, every MCP client using the old key stops working until you update its config. Update the URL or the `Authorization` header in your client config and reconnect.

### Self-hosted: how do I expose the MCP endpoint?

The MCP server starts as part of the Postiz backend and is reachable at `/mcp` (Bearer auth), `/mcp/:apiKey` (key in URL), and `/mcp-oauth` (OAuth-protected). Your reverse proxy must forward these paths to the backend and support streaming HTTP (`Transfer-Encoding: chunked`). See [Reverse Proxies](/self-host/reverse-proxies/caddy).

### Can MCP read or reply to comments?

Not today. The current tool set is read-only on integrations and write-only on posts/media. There is no `getComments` or `replyToComment` exposed via MCP. Comment replies must be triggered through the Postiz UI.