Skip to main content
Glama
Adobe Target MCP

Adobe Target MCP

Adobe Target MCP

npm version npm downloads Node License

A Model Context Protocol server for Adobe Target. Lets AI assistants manage A/B and Experience Targeting activities through the Adobe Target Admin API, with automatic Adobe IMS OAuth token handling (fetch, cache, refresh).

Quickstart

Via Claude Code:

claude mcp add adobe-target \
  -e ADOBE_CLIENT_ID=your-client-id \
  -e ADOBE_API_KEY=your-client-id \
  -e ADOBE_CLIENT_SECRET=your-client-secret \
  -e ADOBE_TENANT=your-tenant \
  -- npx -y @upendra_sengar/adobe-target-mcp

Related MCP server: hackle-mcp

Requirements

  • Node.js >= 18

  • An Adobe Developer Console project with:

    • The Adobe Target API enabled

    • OAuth Server-to-Server credentials (client ID + client secret)

  • Your Adobe Target tenant name (the subdomain in mc.adobe.io/<tenant>/target/...)

Configuration

All configuration is passed as environment variables — no keys are ever stored in this package.

Variable

Required

Description

ADOBE_CLIENT_ID

✅

Client ID from your Adobe Developer Console project

ADOBE_API_KEY

✅

API key sent as X-Api-Key (same value as the client ID)

ADOBE_CLIENT_SECRET

✅

Client secret from your Adobe Developer Console project

ADOBE_TENANT

✅

Adobe Target tenant subdomain

PORT

—

Port for SSE / Streamable HTTP modes (default 3001)

MCP_TOOL_LOGGING

—

true to enable verbose per-tool logging

MCP_LOG_FILE

—

File path for logs in stdio mode

For local development you can copy .env.example to .env instead.

Installation

These clients all use the same mcpServers format:

{
  "mcpServers": {
    "adobe-target": {
      "command": "npx",
      "args": ["-y", "@upendra_sengar/adobe-target-mcp"],
      "env": {
        "ADOBE_CLIENT_ID": "your-client-id",
        "ADOBE_API_KEY": "your-client-id",
        "ADOBE_CLIENT_SECRET": "your-client-secret",
        "ADOBE_TENANT": "your-tenant"
      }
    }
  }
}

Add to your workspace .vscode/mcp.json:

{
  "servers": {
    "adobe-target": {
      "command": "npx",
      "args": ["-y", "@upendra_sengar/adobe-target-mcp"],
      "env": {
        "ADOBE_CLIENT_ID": "your-client-id",
        "ADOBE_API_KEY": "your-client-id",
        "ADOBE_CLIENT_SECRET": "your-client-secret",
        "ADOBE_TENANT": "your-tenant"
      }
    }
  }
}

Transports

Mode

Command

Use case

stdio (default)

npx -y @upendra_sengar/adobe-target-mcp

Local MCP clients (Claude Desktop, Cursor, …)

Streamable HTTP

npx -y @upendra_sengar/adobe-target-mcp --streamable-http

Remote / shared deployments (POST /mcp)

SSE

npx -y @upendra_sengar/adobe-target-mcp --sse

Legacy SSE clients (GET /sse, POST /messages)

HTTP modes listen on PORT (default 3001).

Tools

Tool

Description

list_activities

List activities with filters (state, name, type, priority) and pagination

get_ab_activity_by_id

Get full details of an A/B activity

create_ab_activity

Create a new A/B activity

update_activity

Update an existing activity

update_activity_name

Rename an activity

update_activity_selector

Update the DOM selector of an activity's experience location

Authentication flow

The server exchanges your client credentials for an Adobe IMS access token (ims-na1.adobelogin.com), caches it, and refreshes it automatically 5 minutes before expiry. Tokens are held in memory only — nothing is written to disk.

Docker

docker build -t adobe-target-mcp .
docker run -i --rm \
  -e ADOBE_CLIENT_ID=your-client-id \
  -e ADOBE_API_KEY=your-client-id \
  -e ADOBE_CLIENT_SECRET=your-client-secret \
  -e ADOBE_TENANT=your-tenant \
  adobe-target-mcp

License

MIT

Available Tools

6 tools
create_ab_activityB

Create an AB activity in Adobe Target.

Inputs: name (required), metrics (required), options (required), locations (required), +3 more.

get_activity_by_idA

Get an Adobe Target activity by ID. Supports both AB and XT activity types. API key and tenant are loaded from environment variables automatically.

Inputs: activityId (required), activityType.

list_activitiesC

List activities from Adobe Target. API key and tenant are loaded from environment variables automatically.

Inputs: name, type, limit, state, +2 more.

update_activityA

Update an Adobe Target activity by injecting new HTML/JS code into a named experience. API key and tenant are loaded from environment variables automatically. Handles: deep-clone of payload, creating new option slots (optionLocalId=0), orphan pruning, offerContent deduplication (prevents UniqueElements AT error), and supports both AB and XT activity types.

Inputs: newCode (required), activityId (required), experienceName (required), activityType.

update_activity_nameA

Update the name of an activity in Adobe Target. API key and tenant are loaded from environment variables automatically.

Inputs: name (required), activityId (required).

update_activity_selectorA

Update one or more location selectors in an Adobe Target activity. Use this when the activity's primary location selector (e.g. "BODY > *:eq(0)") does not match the intended CSS injection target (e.g. "#business"). Returns a list of available location names if the specified name is not found. API key and tenant are loaded from environment variables automatically. Supports both AB and XT activity types.

Inputs: activityId (required), selectorUpdates (required), activityType.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.7/5.0

    Scored across 6 tools

    Disambiguation5/5

    Each tool targets a distinct action: listing, getting by ID, creating AB activities, updating the name, injecting HTML/JS code, or modifying selectors. No two tools overlap in purpose.

    Naming Consistency5/5

    All tools follow a consistent verb_noun pattern in snake_case (e.g., list_activities, update_activity_name, create_ab_activity). No mixing of conventions.

    Tool Count5/5

    With 6 tools, the set is well-scoped for an Adobe Target MCP server. It provides core operations without being excessive or sparse.

    Completeness3/5

    The set covers create, read, and update operations, but notably lacks a delete activity tool. This creates a gap in basic lifecycle management.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      A Model Context Protocol server that enables interaction with the Tradovate API for managing trading contracts, positions, orders, and accounts.
      1
      MIT
    • A
      license
      B
      quality
      D
      maintenance
      A Model Context Protocol server for Hackle API providing tools and resources for querying A/B Test data.
      19
      43 npm
      3
      MIT
    • A
      license
      B
      quality
      D
      maintenance
      A Model Context Protocol server that integrates with the Harvest API v2, enabling time tracking management including listing, creating, updating, and deleting time entries, as well as managing projects, tasks, users and generating reports.
      17
      28 npm
      MIT