DocumentationTargetTarget Guide

[Beta]{class="badge informative"}

Adobe Target MCP server target-mcp

Last update: May 12, 2026

CREATED FOR:

  • Beginner
  • Intermediate
  • User
  • Developer

The Adobe Target MCP integration lets you inspect and analyze A/B tests, personalization activities, and Recommendations criteria directly from your AI assistant. Turn Target’s experimentation and personalization data into plain-language workflows — audit your experiment portfolio, review performance reports, and explore audiences and offers without navigating the UI or writing API calls.

AVAILABILITY
The Adobe Target MCP server is available to all customers in Public Beta. It is currently supported in Claude Web, Claude Desktop, Claude Code, Cursor, and ChatGPT. Support for additional MCP-compatible applications will be added in future releases.

What is the Model Context Protocol? mcp-overview

Marketing and optimization teams increasingly rely on chat-based applications and developer tools — such as Anthropic Claude, OpenAI ChatGPT, Cursor, and Microsoft Copilot Studio — to streamline their day-to-day work. These applications support the Model Context Protocol (MCP), an open standard that lets applications expose back-end tools to large language models (LLMs) in a uniform way.

Adobe Target now provides an MCP server that surfaces experimentation, personalization, and recommendations operations directly inside any MCP-compatible application. Adobe Target acts as the decisioning and execution layer while the AI assistant handles reasoning and explanation — giving teams faster access to optimization insights without navigating multiple product screens or writing queries against the Adobe Target REST API.

IMPORTANT
The Model Context Protocol (MCP) is an emerging open-source standard and may present security or reliability risks. Adobe MCP server integrations and related documentation are provided “as is,” without warranties of any kind.
Connecting MCP clients or servers to Adobe products is a customer-elected configuration, and customers are responsible for evaluating the security and suitability of any MCP integration. Adobe is not responsible for issues arising from misconfiguration, misuse of the MCP, vulnerabilities in third-party implementations, or unintended actions performed through MCP-enabled workflows.
To reduce risk, Adobe encourages testing integrations in a sandbox environment prior to productive use and carefully reviewing and validating all MCP-initiated actions and responses before confirming or relying on them.

Key capabilities mcp-capabilities

The Adobe Target MCP server provides read access to activities, audiences, offers, recommendations, and implementation configuration. With the integration, you can:

  • Inspect and audit experiments - Get status, performance, change history, and QA preview links for any activity without navigating the UI.
  • Analyze results - Retrieve performance, revenue, and A4T reports for A/B, XT, AP, and Auto-Target activities.
  • Explore activities - List, inspect, and analyze A/B and XT activities.
  • Explore audiences and offers - List and inspect audiences, HTML offers, and JSON offers.
  • Explore Recommendations criteria - List and inspect criteria and cart-based algorithms.
  • Audit implementation - Review at.js settings, response tokens, and per-entity revision history.
NOTE
Write tools (create, update, activate, deactivate) are not exposed via the public MCP catalog in Public Beta. All currently available tools are read-only. Write access will be available in a future release.

The Adobe Target MCP server exposes 23 read-only tools across 10 categories — from activity inspection and reporting to audience exploration and QA previews. For the complete parameter reference, see MCP server tools reference.

To explore what you can do with the Adobe Target MCP server — including step-by-step prompt walkthroughs — see Use cases and walkthroughs.

To connect the Adobe Target MCP server to your AI assistant — including prerequisites, client-specific configuration, and troubleshooting — see Get started.

Frequently asked questions mcp-faq

Which MCP clients are supported?
The Adobe Target MCP server is currently available for Claude Web, Claude Desktop, Claude Code, Cursor, and ChatGPT. Support for additional MCP-compatible applications may be added in future releases.
What Adobe Target objects can I access via MCP?
You can access activities (A/B, XT, AP), audiences, offers, properties, mboxes, Recommendations criteria, response tokens, at.js configuration, A4T reports, and entity revision history. All 23 currently available tools are read-only.
Can the MCP server create or modify activities?
Not in Public Beta. The public MCP catalog currently exposes 23 read-only tools. Write operations (create, update, activate, deactivate) are not yet available via the public MCP server. Write access will be available in a future release.
Do I need developer access to use the MCP server?
No. The MCP server is designed for both marketing and technical personas. Marketers can interact with it using natural language prompts in any supported MCP client, while developers can use it in tools like Claude Code or Cursor.
Is my Adobe Target data sent to the MCP client provider?
When you submit a prompt, the MCP client may send relevant context (including Adobe Target data returned by the MCP server) to its model for processing. Review the privacy and data-handling policies of your MCP client provider before connecting to production data. Adobe’s data handling is governed by the Adobe Privacy Policy and Data Protection Terms.
Can write operations cause unintended changes to live activities?
Write tools are not available via the public MCP catalog in Public Beta — all 23 currently exposed tools are read-only. When write tools are introduced in a future release, they will include safety annotations and confirmation gates so that no state-changing action is executed without explicit user confirmation.
What permissions do I need in Adobe Target?
Observer role or higher grants access to all 23 read-only tools available in Public Beta. Write tools are not yet exposed via the public MCP catalog, so Editor and Approver role privileges do not unlock additional MCP tools at this time. Contact your Adobe Target administrator if you are unsure about your current access level.
Can I use the MCP server across multiple Target organizations or properties?
The MCP server scopes operations to the organization associated with your authenticated Adobe IMS credentials. If you have access to multiple properties within that organization, you can query by property using the list_target_properties tool and filter subsequent requests accordingly.
recommendation-more-help
target-help-main