Skip to main content
·4 min read·Lucas Keeler

Ask your AI which campaign actually worked: the Persequor MCP server is live

Persequor now speaks MCP. Connect Claude (or any MCP client) to your workspace and ask plain-English questions about real attribution data — then have it create and update leads, pull the agency rollup, and more. Here's what it does and how to turn it on.

You already talk to an AI assistant all day. Now it can talk to your attribution data.

Persequor ships a Model Context Protocol (MCP) server — a standard, secure way for AI clients like Claude to read from and act inside your Persequor workspace. Connect it once and "what's my best-performing campaign this month?" becomes a question you ask in chat, answered from your real numbers, not a guess.

What MCP actually is

MCP is an open standard for letting an AI assistant use external tools safely. Instead of copying numbers out of a dashboard and pasting them into a chat window, the assistant calls Persequor directly — with your permission, scoped to your account — and works with live data. Think of it as giving your AI a set of well-labeled buttons it's allowed to press.

What you can ask

The server exposes eleven tools covering the data you live in:

  • Read your pipeline — list and inspect leads, orders, campaigns, and workspaces.
  • Understand attribution — pull a single lead's full touchpoint journey, a workspace's blended metrics, or campaign-level performance across attribution models.
  • See the whole agency — the agency rollup, across every client workspace at once.
  • Take action — create and update leads without leaving the conversation.

So instead of ten browser tabs, you ask: "Across all my clients, which workspace had the best ROAS last week, and which campaign drove it?" — and your assistant answers from the same source of truth your dashboard uses.

Built for agencies first

If you run multiple client workspaces, this is where it earns its keep. An agency API key reaches every workspace you manage. One question spans the whole book of business. No switching, no per-client logins, no exporting five CSVs to compare them.

Read and write — with real guardrails

Letting an AI take action only works if the rails are solid. Ours are:

  • Your permissions, enforced. The server respects the same access your dashboard does — a key can only see what you can see.
  • Every write is audit-logged. Creates and updates are recorded, so there's always a trail.
  • Demo and read-only workspaces stay read-only. No accidental writes where they don't belong.

How to turn it on

It takes about a minute:

  • Generate an API key in Settings → API. (Agencies: use an agency-scoped key to reach all workspaces.)
  • Add the server to your MCP client. In Claude Code that's one command:

claude mcp add --transport http persequor https://www.persequor.ai/api/mcp --header "Authorization: Bearer pk_your_key"

  • Start asking. Your assistant discovers the tools automatically and uses them as needed.
The dashboard answers the questions you remember to ask. An assistant wired into your data answers the ones you didn't.

Full endpoint and tool details live in the developer docs. Questions or want a hand wiring it up? hello@persequor.ai.