Use paperplane in Lindy

Lindy agents can use tools from any MCP server, and paperplane’s MCP endpoint is public — no API key, no OAuth. That means a Lindy that handles your ops inbox, chases invoices, or runs a compliance workflow can also put real paper in the USPS mail stream. Honest framing up front: this is user-side setup. paperplane has no native listing in Lindy’s integration library yet — you are adding our standard MCP server to your own workspace, the same endpoint every other MCP client uses.

Connect it

• In Lindy, open Integrations and click + Add, then choose MCP.
• In the “Connect MCP server” dialog, paste the Server URL: https://sendpaperplane.com/api/mcp and give it a display name like paperplane.
• Skip the Advanced credentials section — the server supports anonymous access, so there is no bearer token or header to add.
• Click Connect. The server appears in your list with its tool count (three), and you can choose whether it is available to just you or workspace-wide.
• In the guardrails settings, consider requiring approval for send_letter while leaving the two read-only tools automatic — see the FAQ below.

What your agent can now do

• quote_letter — free, read-only. Returns the price (First-Class $1.99, Certified $12.99, Certified + electronic Return Receipt $14.99, Priority $24.99) and, when given the real recipient and letter content, a single-use confirmation_token bound to exactly that letter at exactly that price.
• send_letter — the write tool. Requires the token, so the recipient, content, and price can’t drift between what was quoted and what gets mailed. Live sends return a Stripe payment link (or draw from a prepaid credit balance); sandbox: true completes instantly and free.
• get_letter_status — read-only order tracking through screening → printing → mailed → delivered, so an agent can report back or branch on delivery.

Sandbox-first walkthrough

Before wiring paperplane into a real workflow, run one sandbox send in a test conversation with your Lindy:

"Using the paperplane tools: quote a one-page First-Class letter to
Maria Alvarez, 1 Main St, Richmond VA 23220 saying 'This is a test of our
mail integration.' Show me the price, then send it in sandbox mode."

You should see the agent call quote_letter, surface a $1.99 quote with a confirmation_token, then call send_letter with the same letter plus sandbox: true. The response is a completed sandbox order with an id you can feed to get_letter_status. Nothing was printed or charged. When the flow looks right, remove the sandbox instruction; for hands-off live sending, fund a prepaid credit balance and give the agent the credit code.

Where this fits in a Lindy workflow

The natural pattern is mail as the last step of a flow Lindy already runs: an overdue-invoice chaser that escalates from email to a certified demand letter, a compliance Lindy that mails termination or notice letters your ERP generates, an onboarding agent that sends a real welcome note. Because quote_letter is free, agents can price options mid-conversation (“certified or first-class?”) before anyone commits to anything.

Prefer plain HTTP from a Lindy step instead of MCP? The same backend is REST — POST /v1/quotes, POST /v1/orders, GET /v1/orders/:id — documented on the developer docs. And the full mechanics of the quote→send token contract are in how AI agents mail letters via MCP, or live in the no-signup demo.