# MCP server (/mcp)



The LettrLabs **MCP server** lets MCP-compatible AI agents — Claude, ChatGPT,
Cursor, and others — work with your LettrLabs account directly from chat. Your
agent can draft campaigns, preview pricing, and submit mailings without you
copying anything into a separate dashboard.

<Callout>
  **Public beta.** The MCP server is live at `https://mcp.lettrlabs.com/mcp` for
  all LettrLabs customers. We're still refining it, so connection flows and
  behavior may change — tell your LettrLabs contact if anything looks off.
</Callout>

## What you need [#what-you-need]

* **A LettrLabs API key.** Generate one in the app and keep it somewhere secure —
  see [Authentication & API keys](/authentication-api-keys). The same key
  authorizes your agent.
* **A plan with API access.** The MCP server calls the LettrLabs public API,
  which is available on plans with OpenAPI integrations enabled. If your agent
  reports that your subscription doesn't include OpenAPI integrations, ask your
  LettrLabs rep to enable it.

## Connect your agent [#connect-your-agent]

The MCP server is available at:

```
https://mcp.lettrlabs.com/mcp
```

The exact menu names vary by agent, but the flow is the same:

1. In your agent, add a **custom MCP connector** (for example, Claude:
   **Settings → Connectors → Add custom connector**; ChatGPT:
   **Settings → Connectors → Add MCP connector**).
2. Set the server **URL** to `https://mcp.lettrlabs.com/mcp`.
3. **Authenticate with your API key.*&#x2A; Agents that use a consent page (such as
   Claude) open a LettrLabs authorization page where you paste your key. Agents
   that use header authentication send it as the &#x2A;*`X-API-KEY`** header.
4. Enable the connector for your conversation, then try asking it to
   &#x2A;"list my recent orders."*

For IDE and CLI agents that take a JSON config, point the server at the same URL
with your key in the `X-API-KEY` header:

```json
{
  "url": "https://mcp.lettrlabs.com/mcp",
  "headers": {
    "X-API-KEY": "your_api_key_here"
  }
}
```

## What you can do [#what-you-can-do]

Once connected, your agent can act on your LettrLabs account on your behalf:

* **Read your account** — list your direct mail orders and their recipients,
  pull printed-piece proofs, and review spend and delivery analytics.
* **Preview pricing** — cost out an order *without* charging anything, so you can
  see the price before you commit.
* **Draft and send campaigns** — start a draft from one of your templates, add a
  mailing list, and submit it for production. Sending is never silent: submitting
  a charge requires an explicit confirmation, so a well-behaved agent shows you
  the price and asks before anything goes out.

## Staying safe [#staying-safe]

* **You authorize each connection.** Your agent connects once, with your
  consent, using the API key you provide.
* **Charges require confirmation.** Nothing is submitted for production until you
  confirm — your agent surfaces the cost first.
* **Treat your key like a password.** It grants access to your account's API
  surface; rotate or revoke it from the same screen if it's ever exposed.