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

# MCP overview

> PostSyncer Model Context Protocol server - same powers as API v1 for AI assistants and autonomous agents.

# PostSyncer MCP

PostSyncer exposes a **Model Context Protocol (MCP)** server so AI assistants and coding agents can manage social content with **tools** (structured actions) instead of ad-hoc HTTP calls.

<Note>
  Two ways to connect: **(1) a connector** — clients that support OAuth (e.g. Claude with custom connectors) add the URL and authorize in PostSyncer, with no token to paste; **(2) a Bearer token** — the same personal access tokens as REST, created in [app.postsyncer.com → Settings → API Integrations](https://app.postsyncer.com/dashboard?action=settings\&section=api-integrations).
</Note>

## Endpoint

| Item                 | Value                                                    |
| -------------------- | -------------------------------------------------------- |
| **MCP URL**          | `https://postsyncer.com/mcp`                             |
| **Auth (connector)** | OAuth — authorize in PostSyncer when prompted (no token) |
| **Auth (token)**     | `Authorization: Bearer YOUR_TOKEN`                       |
| **Transport**        | Streamable HTTP                                          |

### Choosing an auth method

* **Connector (OAuth)** — easiest for assistants that support it (Claude: **Settings → Connectors → Add custom connector**). You add `https://postsyncer.com/mcp`, get sent to PostSyncer to click **Allow**, and the client is connected — nothing to copy or rotate. Access is scoped to the workspaces you belong to.
* **Bearer token** — best for config-file clients (Claude Desktop, Claude Code, Cursor) and scripts. Create a personal access token and send it in the `Authorization` header.

See [Client setup](/mcp/integrating-clients) for both.

<Warning>
  Treat tokens like production passwords. They can publish posts, delete content, and disconnect accounts when matching tools are invoked. Connector access can be revoked anytime from your PostSyncer account.
</Warning>

## What the server does

The **PostSyncer** MCP server (name `PostSyncer`, version `1.0.0`) implements tools that mirror API v1:

* **Workspaces** - discover context and connected accounts.
* **Accounts** - list connected social profiles; disconnect when needed.
* **Labels & campaigns** - organize content.
* **Posts** - create, list, read (by id, public URL, or platform post id), analyze public X/Twitter threads with AI, update, delete, schedule.
* **Comments** - list, sync from platforms, create replies, update, delete, hide.
* **Analytics** - summary, workspace, post, and account metrics; sync post analytics.

Server instructions remind models to start with `list-workspaces` and `list-accounts`, respect workspace boundaries, and only run `delete-*` tools when the user clearly asked.

## Relationship to the REST API

* **Parity**: MCP tools call the same backend rules as `https://postsyncer.com/api/v1`.
* **Choose MCP** when your client supports Streamable HTTP and Bearer headers (see [Client setup](/mcp/integrating-clients)).
* **Choose REST** for servers, webhooks, mobile apps, or languages without MCP.

## More on PostSyncer.com

* [AI agents & MCP](https://postsyncer.com/agents)
* [OpenClaw + PostSyncer](https://postsyncer.com/openclaw)
* [MCP for assistants (Claude, Cursor, …)](https://postsyncer.com/ai-mcp)

## Next steps

<CardGroup cols={2}>
  <Card title="Client setup" icon="laptop" href="/mcp/integrating-clients">
    Claude Desktop, Claude Code, Cursor, and other Streamable HTTP clients.
  </Card>

  <Card title="Tools reference" icon="wrench" href="/mcp/tools-reference">
    Full list of MCP tool names grouped by domain.
  </Card>
</CardGroup>