Skip to content
Book a demo

Connect documentation

All documentation
Documentation

Using Adverity MCP

On this page

This guide explains what you can do with Adverity MCP and the tools available.

Introduction

What is Adverity MCP?

Adverity Model Context Protocol (MCP) connects your AI assistant to Adverity, allowing you to monitor, configure, and control your data pipelines using natural language. Instead of navigating the Adverity interface, you can ask your AI tool to check authorization status, troubleshoot failed fetches, set up new datastreams, or download data extracts on demand.

What are the benefits of using Adverity MCP?

With Adverity MCP you can:

  • Monitor your data pipelines without leaving your AI assistant.

  • Get instant answers about datastream status, fetch history, and authorization status.

  • Download data extracts on demand through simple prompts.

  • Troubleshoot failed jobs by asking questions in plain language.

What can I do with Adverity MCP?

With Adverity MCP you can monitor, configure, and control your data pipelines:

  • Create and manage datastreams, their configurations, and current status.

  • Create and manage authorizations and check whether each connection is currently authorized.

  • Configure which accounts an authorization is permitted to collect from.

  • Trigger fetches, review job history, and troubleshoot errors.

  • Download data extracts from successful datastream executions.

  • Create, update, and explore workspaces and their storage.

  • Create and manage transformations as part of datastream setup.

  • Manage mapping table entries for data harmonization.

  • Map datastream source columns to target columns.

  • Create and manage destinations and which datastreams load into them.

  • Tag datastreams and find datastreams by tag.

  • Switch a datastream between scheduled and smart pulling.

  • Query the data Adverity has loaded into a workspace’s warehouse.

  • List and manage instance users.

  • Manage email, Slack, Microsoft Teams, and webhook notification subscriptions.

  • Review data-quality monitors and manage which datastreams they watch.

  • Look up answers from the official Adverity documentation.

Is my data secure?

When you sign in with Adverity, you select the workspace and the permissions the AI assistant gets, up to the permissions you have when you sign in, and the connection expires after 30 days. When you connect with your own Connect API key, the AI assistant only has access to data permitted by your key’s permissions. All communication happens through secure, authenticated connections.

How do I connect my AI tool to Adverity MCP?

For connection instructions, see Setting up Adverity MCP.

Can I use these tools without my own AI assistant?

Yes, where Connect Agent is available for your account. Connect Agent is an AI assistant built into Adverity that uses the Adverity MCP tools, with no setup. For more information, see Using Connect Agent.

Available MCP tools and use cases

The MCP server provides 41 tools organized by domain. For the full list, see Adverity MCP tools. For example tasks and prompts, see Adverity MCP use cases.

How the tools are named

For how tool names show what a tool does to your data, see Adverity MCP tools.

Confirmation for write actions

Every tool action that creates, changes, or deletes data is two-phase. The confirm parameter defaults to false, and the two phases are split between your AI assistant and you:

  1. The assistant calls the action with confirm: false (the default) and receives confirmation_required: true together with a preview of the change. Fields such as will_create, will_update, or warning describe the impact of the change. Nothing is written at this point.

  2. The assistant shows you the preview. After you approve it, the assistant calls the same action with confirm: true to apply the change.

You never set confirm yourself. It is the assistant’s signal that a human has approved this specific change, and the tool descriptions instruct the assistant to obtain that approval first. If you script against MCP without an AI assistant, your code follows the same contract: call, inspect the preview, call again with confirm: true. Triggering fetches (trigger, trigger_fixed, stop_job, retry_job in fetch_write) and updating OAuth tokens or refreshing metadata (update_token, refresh_metadata in authorization_write) follow the same two-phase pattern. The one exception is authorize in authorization_write, which returns the OAuth URL for a connection straight away, without a preview.

For the tools that use this pattern, see Adverity MCP tools.

Use cases

For example tasks and prompts, see Adverity MCP use cases.