Patto

Remote MCP server for PR, link building and sales outreach: manage contacts, campaigns, email drafts, replies and follow-ups from your own AI assistant.

Hosted MCP Server

npx add-mcp 'https://pattorelay.com/mcp'

Installs into Claude Code, Codex, Cursor and more

Documentation

Give your AI the tools to put outreach into action.

Turn a plan into a campaign.

Your assistant develops the approach. Patto stores the campaign brief and links the prospects and drafts, giving your assistant a shared structure for the work from first contact to follow-up.

Start each campaign from the last one.

Patto stores the brief, links prospects and sent messages, and records who replied and who was won — so the next campaign starts from what the last one sent and what came back.

Every reply answered in its thread.

Replies are read as they arrive and answered as drafts inside their thread, and Slack or email tells you when one is waiting for release. Released drafts leave through your brand’s own mail account, so the whole exchange stays in your mailbox.

Keep contacts and next steps in reach.

Patto stores organizations, contacts, prospect stages, follow-up dates and recorded interactions. Your assistant can retrieve that working context in a later conversation and help you decide what to do next.

Coordinate outreach across your brands.

Each brand has its own sender identity, mailbox and campaigns. Contacts and organizations are shared across brands; an opt-out applies to the brand it answers. Multi-brand management is available on Standard.

What your assistant does with it

Brands

Several sender identities, each with its own mailbox and its own signature.

Organizations and contacts

The people you deal with, with their roles, their topics and the history of what was sent.

Prospects and follow-ups

A pipeline with stages and dates, and an answer to what is due today.

Campaigns

Outreach grouped under a campaign and kept consistent across its messages. Every campaign shows how many were contacted, replied and won, and what was sent to each.

Mailbox

Replies read and answered as drafts in their thread, with a Slack or email notice for each one waiting for release.

Interactions

What happened — a call, a reply, a meeting — recorded where the next message can use it.

Example use cases

Link building

Get resource pages, city portals and blogs to link to what you published.

PR and media relations

Pitch a story to the journalists who cover it, from your own mailbox.

Sales prospecting

Find the companies that fit, approach them and keep the pipeline moving.

From the first search to the follow-up

Example conversation with example data.

Your assistant Patto connected

Find organizations that could use the new product line. Check each against our criteria and keep the ones that fit.

Your assistant

Web search ×14

Patto

organizations_search ×14 interactions_list ×5 organizations_create ×5 contacts_create ×5 prospects_create ×5

I looked at fourteen organizations. Five fit: each has expanded recently, has a named contact for the topic and has not heard from any of your brands yet. They are saved as qualified prospects, ranked A to C, with the reason and the source page in each record. The other nine are not saved.

Set up a campaign for the autumn launch with these five, building on what worked in spring, and draft a first message to each for my review.

Patto

campaigns_list campaigns_get prospects_list mail_drafts_list interactions_list campaigns_create prospects_update ×5 mail_draft ×5

In spring, 11 of 40 replied — most of them to the messages that led with the delivery date, few to the ones about price. The autumn campaign leads with the date; goal, angle and audience are saved. The five prospects are under it, each with a follow-up in a week, and every message builds on that prospect’s reason to fit. All five wait in your queue.

You review and release the five drafts. Two days later:

Two replies arrive. They are read and answered automatically.

Patto

mail_search mail_get_thread ×2 mail_draft_reply ×2 interactions_record prospects_update ×2

One asks for more material, one wants a call on Tuesday. Both answers are drafted inside their threads, the call is recorded, and both prospects are marked as replied.

Slack

A mail draft is waiting for approval.

Brand: Your brand

Subject: Re: Material for the autumn launch

Review draft Slack

A mail draft is waiting for approval.

Brand: Your brand

Subject: Re: A call about the autumn launch

Review draft

You release both answers from the Slack link. A week after the first messages:

What is due today? Draft the follow-ups.

Patto

prospects_due_followups mail_draft ×3

The three who have not answered yet. Their follow-ups are drafted and wait in your queue.

Bring your own AI

Connect Claude, ChatGPT or another assistant you already use through MCP. It brings the research, reasoning and writing; Patto provides the structured workspace and the tools to act on that work.

Not another AI bill.

Patto charges for the workspace and infrastructure, not for generating text. Your AI provider’s own pricing and usage limits still apply.

Work with your chosen assistant.

Patto connects to an MCP-compatible assistant instead of supplying its own writing model. You choose the assistant and the model options it makes available.

Written by your assistant, not by a persona.

Nothing to upload, no persona to maintain. The assistant that writes is the one you already briefed, with the context and tools it has in the conversation.

Keep the workspace when the conversation changes.

Contacts, campaigns and recorded activity stay in Patto. A connected assistant can retrieve them in a new conversation, without relying on the old chat to hold the working record.

Your AI develops the approach. Patto provides the structure, connects the mailbox and enforces the rules when actions are carried out.

Give your AI responsibility. Set the boundaries.

Choose who releases each draft.

Every outgoing message starts as a draft. Review and release it yourself, or enable assistant release on a paid plan. Patto checks the sending rules either way.

An opt-out is final for its brand.

An objection blocks every further message from that brand, and no request can override it. An address that bounces is blocked for all your brands. A draft that lacks the objection notice is flagged before it goes anywhere.

Limits that protect the recipient.

Patto checks recipient cooldowns and sending limits before release. Replies are exempt from the first-contact cooldown. Retrying the same send operation does not send it again.

Everything is on the record.

Every action of the assistant and every message sent is logged with the user and the time. That is what makes the access defensible — to your client, your management and yourself.

Access you can take back.

Mailbox passwords and Microsoft 365 tokens are stored encrypted and never appear in a result, a log or a form. The assistant’s tokens are short-lived and bound to your account and the areas you granted, and the service never fetches a URL of its own. How Patto handles mailbox access

Hosted in Germany, under the GDPR.

A German company on servers in Germany. The data processing agreement pursuant to Article 28 GDPR comes with every plan.

Connect your AI to a working outreach setup.

  1. Create your account. Email address, password, activation link.
  2. Add a brand and its mailbox. Sender name and address, then IMAP and SMTP, Connect with Microsoft for a Microsoft 365 mailbox, or a managed mailbox from Patto. Passwords and tokens are stored encrypted and never appear in a result, a log or a form.
  3. Connect your assistant. Add the connector URL in Claude, ChatGPT or your MCP client of choice, sign in here, approve the access. How to connect Claude or ChatGPT
  4. Work, then release. Tell your assistant about your brand and what you want to achieve. It plans the campaign, finds the contacts that fit it and drafts your messages. Slack or email tells you when a draft is waiting; release it in Patto, or let the assistant release it once you have enabled that on a paid plan.

Plans

Your AI, your tokens — nothing metered here

Every plan speaks the same connector and gives your assistant the same tools. What you pay for here is the workbench — the assistant is the one you already have.

All you need

An assistant with MCP connectors.

Claude, ChatGPT on a paid plan with Developer Mode, or another AI/LLM assistant with MCP support.

A mailbox you send from.

IMAP and SMTP, Microsoft 365 through Connect with Microsoft, or a managed mailbox from Patto. Its access is stored encrypted and never shown again.