adobe-cja-mcp

adobe-cja-mcp

MCP server for Adobe Customer Journey Analytics, enabling AI-powered analytics queries including reports, breakdowns, trends, and dimension searches through Claude and other MCP clients.

Category
访问服务器

README

Adobe Customer Journey Analytics MCP Server

PyPI version Python 3.10+ License: MIT

Model Context Protocol (MCP) server for Adobe Customer Journey Analytics (CJA), enabling AI-powered analytics queries through Claude and other MCP clients.

⚡ Quick Start (Simplest Setup)

No repository cloning required! Users can run this MCP server with just 2 steps:

1. Install uv (Python package manager)

# macOS/Linux
brew install uv

# Or with curl
curl -LsSf https://astral.sh/uv/install.sh | sh

# Windows
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

2. Configure Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "Adobe CJA": {
      "command": "uvx",
      "args": ["adobe-cja-mcp"],
      "env": {
        "ADOBE_CLIENT_ID": "your_client_id_here",
        "ADOBE_CLIENT_SECRET": "your_client_secret_here",
        "ADOBE_ORG_ID": "your_org_id@AdobeOrg",
        "ADOBE_DATA_VIEW_ID": "dv_your_dataview_id"
      }
    }
  }
}

That's it!

Restart Claude Desktop and uvx will automatically download and run the MCP server from PyPI - just like npx for Node.js!


Overview

This MCP server provides tools for querying Adobe CJA data, including:

  • Running analytics reports with dimensions and metrics
  • Multi-dimensional breakdown analysis
  • Time-series trend analysis at various granularities
  • Searching and filtering dimension values
  • Listing available dimensions and metrics
  • Data view configuration access

Features

Core Reporting Tools (MVP):

  • Run ranked reports with dimensions and metrics
  • Get top N items for any dimension
  • List available dimensions and metrics
  • Access data view configuration

Advanced Reporting:

  • Breakdown Reports: Multi-dimensional analysis (e.g., pages by device type)
  • Trended Reports: Time-series analysis with hourly/daily/weekly/monthly granularity
  • Dimension Search: Find specific dimension values (pages, products, campaigns)
  • Segment Support: Filter reports with segment IDs

Adobe CJA API Permissions

Required Permissions Scope

For this MCP server to function, your Adobe API credentials (OAuth 2.0 Server-to-Server) must have the following permissions granted in Adobe Admin Console.

Authentication Scopes

Your API credentials require these OAuth scopes:

openid
AdobeID
read_organizations
additional_info.projectedProductContext
cja.reporting         # CRITICAL - Required for all reporting operations
cja.workspace         # Required for workspace objects (projects, filters, etc.)

Critical Permissions (Required for Core Functionality)

Without these permissions, the MCP server cannot perform analytics queries:

1. CJA Reporting API - Ranked Reports

  • API Endpoint: POST https://cja.adobe.io/reports
  • Permission: CJA Reporting API Access
  • Required For:
    • Running analytical reports with dimensions and metrics
    • Querying website performance data (sessions, page views, conversions)
    • Generating attribution analysis
    • Creating breakdown reports
  • MCP Tools Blocked Without Permission:
    • cja_run_report - Main reporting tool
    • cja_get_top_items - Top performers analysis
    • cja_get_trended_report - Time-series analysis
    • cja_get_breakdown_report - Multi-dimensional breakdowns
    • cja_get_sessions_data - Session analytics
    • cja_get_conversions_data - Conversion tracking
    • cja_get_attribution_analysis - Attribution modeling
    • cja_get_funnel_analysis - Funnel analysis

2. CJA Reporting API - Top Items

  • API Endpoint: GET https://cja.adobe.io/reports/topItems
  • Permission: CJA Reporting API Access
  • Required For:
    • Ranking dimension items by metrics
    • Finding top pages, products, campaigns
  • MCP Tools Blocked Without Permission:
    • cja_get_top_items (alternative implementation)

Working Permissions (Read-Only Access)

These permissions are typically granted by default for read-only CJA API access and should already be available:

  1. GET /data/dataviews/{id}/dimensions - List available dimensions
  2. GET /data/dataviews/{id}/metrics - List available metrics
  3. GET /data/dataviews/{id} - Get data view configuration details
  4. GET /calculatedmetrics - List calculated metrics
  5. GET /filters - List filters/segments
  6. GET /dateranges - List date ranges
  7. GET /annotations - List annotations
  8. GET /projects - List Analysis Workspace projects
  9. GET /data/connections - List data connections

Optional Permissions (Enhanced Features)

These permissions enable additional features but are not required for basic operation:

Individual Dimension/Metric Details

  • Endpoints:
    • GET /data/dataviews/{id}/dimensions/{dimId}
    • GET /data/dataviews/{id}/metrics/{metricId}
  • Benefit: Get detailed metadata for specific dimensions/metrics
  • Workaround: Use list endpoints instead

Filter Validation

  • Endpoint: POST /filters/validate
  • Benefit: Validate filter definitions before use
  • Workaround: Test filters directly in reports

List All Data Views

  • Endpoint: GET /dataviews
  • Benefit: Discover all available data views
  • Workaround: Use configured data view ID from environment variables

Setting Up Permissions in Adobe Admin Console

Step 1: Navigate to API Credentials

  1. Log in to Adobe Admin Console
  2. Navigate to ProductsCustomer Journey Analytics
  3. Click on API Credentials
  4. Select your OAuth Server-to-Server credential (Client ID)

Step 2: Add Product Profile

  1. Click Add Product Profile
  2. Select a profile that includes CJA Reporting API Access
  3. Ensure the profile has the following permissions:
    • Reporting API: Full access to POST /reports and GET /reports/topItems
    • Workspace API: Access to filters, calculated metrics, projects
    • Data Views: Read access to configured data views

Step 3: Verify OAuth Scopes

In the API credential configuration, verify these scopes are enabled:

  • openid
  • AdobeID
  • read_organizations
  • additional_info.projectedProductContext
  • cja.reportingCRITICAL
  • cja.workspaceCRITICAL

Step 4: Generate New Credentials (if needed)

If updating an existing credential doesn't enable reporting access:

  1. Create a new OAuth Server-to-Server credential
  2. Add CJA API as a service
  3. Select a Product Profile with full Reporting API access
  4. Copy the new Client ID and Client Secret to your .env file

Example Queries and Expected Results

Session Analysis

Query:

Show me session count in a bar chart for the last 30 days

What Happens:

  1. Claude identifies you want session metrics over time
  2. Calls cja_get_trended_report or cja_run_report with:
    • Metric: sessions or visits
    • Date range: Last 30 days
    • Granularity: day
  3. Formats results as a text-based bar chart or table
  4. Returns daily session counts with visualization

Expected Output:

Session Count - Last 30 Days

Oct 1  ████████████████ 1,245
Oct 2  ██████████████ 1,108
Oct 3  ███████████████ 1,189
...
Oct 30 ████████████████████ 1,523

Total Sessions: 38,420
Average: 1,281 sessions/day
Peak: Oct 30 (1,523 sessions)

Top Pages Analysis

Query:

What are the top 10 pages by page views this month?

What Happens:

  1. Calls cja_get_top_items with:
    • Dimension: page or pageName
    • Metric: pageviews
    • Date range: This month
    • Limit: 10
  2. Returns ranked list of pages

Expected Output:

Top 10 Pages by Page Views - October 2025

1. /home                    15,234 views
2. /products                 8,901 views
3. /about                    6,543 views
4. /contact                  4,321 views
5. /blog/article-123         3,987 views
6. /pricing                  3,456 views
7. /features                 2,890 views
8. /blog                     2,543 views
9. /documentation            2,234 views
10. /support                 1,987 views

Total: 52,096 page views

Conversion Funnel

Query:

Show me the checkout funnel conversion rates for last week

What Happens:

  1. Calls cja_get_funnel_analysis with predefined checkout steps
  2. Or calls cja_run_report multiple times for each funnel step
  3. Calculates drop-off rates between steps

Expected Output:

Checkout Funnel - Last 7 Days

Step 1: Product Page     → 10,000 visitors (100.0%)
         ↓ 45.0%
Step 2: Add to Cart      →  4,500 visitors ( 45.0%)
         ↓ 66.7%
Step 3: Checkout Started →  3,000 visitors ( 30.0%)
         ↓ 50.0%
Step 4: Payment Info     →  1,500 visitors ( 15.0%)
         ↓ 80.0%
Step 5: Purchase         →  1,200 visitors ( 12.0%)

Overall Conversion Rate: 12.0%
Biggest Drop-off: Product Page → Add to Cart (55%)

Marketing Attribution

Query:

Show me first-touch attribution for conversions in the last 14 days

What Happens:

  1. Calls cja_get_attribution_analysis with:
    • Attribution model: first_touch
    • Success event: orders or conversions
    • Date range: Last 14 days
  2. Returns marketing channel credit distribution

Expected Output:

First-Touch Attribution - Last 14 Days
Total Conversions: 450

Channel              Conversions    % of Total    Revenue
──────────────────────────────────────────────────────────
Organic Search            180         40.0%      $45,000
Paid Search               135         30.0%      $33,750
Direct                     68         15.1%      $17,000
Email                      45         10.0%      $11,250
Social Media               22          4.9%       $5,500

Top First-Touch Channel: Organic Search
Highest Revenue/Conversion: Paid Search ($250 avg)

Available MCP Tools

Core Reporting

  • cja_run_report - Run custom analytics reports
  • cja_get_top_items - Get top-performing items for a dimension
  • cja_get_trended_report - Get time-series trends
  • cja_get_breakdown_report - Get multi-dimensional breakdowns

Metadata

  • cja_list_dimensions - List available dimensions
  • cja_list_metrics - List available metrics
  • cja_list_calculated_metrics - List calculated metrics
  • cja_list_filters - List filters/segments
  • cja_list_date_ranges - List predefined date ranges
  • cja_get_dataview_info - Get data view configuration

Advanced Analytics

  • cja_search_dimension_items - Search for dimension values
  • cja_get_sessions_data - Analyze session metrics
  • cja_get_conversions_data - Analyze conversion events
  • cja_get_attribution_analysis - Run attribution models
  • cja_get_funnel_analysis - Analyze conversion funnels

License

MIT License - see LICENSE file for details

Support

For issues or questions:

  • Adobe CJA API Documentation: https://developer.adobe.com/cja-apis/docs/
  • Adobe Developer Console: https://developer.adobe.com/console
  • Model Context Protocol: https://modelcontextprotocol.io/

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选