Get Started

How MCP Works

MCP — the Model Context Protocol — is an open standard that lets an AI client call external tools during a conversation. MCP Ads implements an MCP server that exposes your ad accounts as callable tools.

The request flow

Nothing about this is magic, and it helps to know the sequence, because it explains both the speed and the limits.

  1. You describe an outcome

    In plain language: an audit, a report, a change. You do not name a tool.

  2. The client picks tools

    Claude reads the list of available tools and their parameters, and decides which ones can answer the request and in what order.

  3. MCP Ads calls the platform

    The server calls the underlying platform API using your connected, scoped OAuth token. Your credentials never pass through the conversation.

  4. The result returns to the chat

    Live data comes back into the same conversation, where Claude can analyse it, chart it, or use it to decide the next call.

Read tools and write tools

The distinction matters more than any other concept here, because it determines what can happen without your say-so.

Read toolsWrite tools
What they doReports, account structure, audience insight, diagnosticsCreate campaigns, adjust budgets, manage keywords and creatives
Share of the surfaceThe large majority of toolsA deliberately smaller set
When they runWhenever they answer your questionOnly when you ask for that specific outcome
Effect on spendNone. They change nothing.Can change spend, which is why new campaigns start paused

What authentication actually looks like

Each platform connection is a standard OAuth grant between you and that platform. MCP Ads holds the resulting token, encrypted, and uses it only to make the calls you trigger.

  • You authorise on the platform's own screen — MCP Ads never sees your platform password
  • Tokens are scoped to the permissions that platform grants, and are revocable per connection at any time
  • The MCP endpoint itself is authenticated too: an unauthenticated request is rejected rather than answered
Note

There is no API key to copy and no client secret to paste. If a guide anywhere tells you to put a secret into a config file for MCP Ads, it is out of date.

Safety boundaries

The design goal is that connecting the server extends what Claude can find out, without quietly extending what it can spend.

  • Every action runs inside an account you explicitly authorised — never one you did not
  • Deleting campaigns and removing budgets outright are not exposed as tools
  • New campaigns are created paused by default, so a misread instruction cannot start spending
  • Any connection can be revoked from your dashboard, which immediately invalidates the token

Common questions

Is MCP specific to Claude?

No. MCP is an open protocol and any client that implements it can connect to an MCP server. Claude was the first widely used client to support it, which is why most documentation — including ours — uses Claude for the examples.

Does Claude see my whole account all the time?

No. Nothing is read until a tool call happens, and a tool call only happens in response to something you asked. There is no background sync and no standing copy of your account.

What stops Claude from calling a write tool by accident?

Write tools are a small, explicit subset, and the destructive operations are simply not exposed. New campaigns start paused. The practical guardrail is that the worst realistic outcome of a misunderstanding is a paused campaign you delete, not unplanned spend.

What is the difference between the read-only and write tools in the reference?

Every tool page in the API reference states which it is in its first line, and write tools note that Claude asks you to confirm before running them.

Ready to connect?

Start free, connect the platforms you use, and run your first workflow from Claude in minutes.