home / mcp / optimizely dxp mcp server

Optimizely DXP MCP Server

Provides a REST API and stdio workflow to manage deployments, logs, storage, and real-time monitoring for Optimizely DXP with automation-ready tooling.

Installation
Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "jaxondigital-optimizely-dxp-mcp": {
      "url": "https://mcp.example.com/mcp",
      "headers": {
        "REDIS_HOST": "localhost",
        "REDIS_PORT": "6379",
        "DXP_MCP_HOST": "0.0.0.0",
        "DXP_MCP_MODE": "http",
        "DXP_MCP_PORT": "3001",
        "REDIS_ENABLED": "false",
        "REDIS_PASSWORD": "your-redis-password",
        "OPTIMIZELY_PROJECT_ID": "your-project-id",
        "OPTIMIZELY_PROJECT_KEY": "your-key",
        "OPTIMIZELY_PROJECT_SECRET": "your-secret"
      }
    }
  }
}

You transform Optimizely DXP into an AI-powered, autonomous operations platform. The MCP Server lets you manage deployments, logs, databases, storage, and real-time monitoring through a REST API or a zero-setup stdio workflow. It enables scalable automation, real-time event subscriptions, and error-resilient operations so you ship faster and with less manual toil.

How to use

You interact with the MCP Server through an MCP client or automation platform. There are two transport modes, and you can choose the one that fits your workflow: a direct HTTP REST API mode for remote automation, or a stdio mode for local AI clients and Claude Desktop. In either mode, you access 45 tools that cover permission checks, deployments, logs, databases, storage, multi-project management, configuration utilities, and advanced features. Use these tools to perform operations, subscribe to real-time deployment events, and retrieve structured data for your automation pipelines.

How to install

Prerequisites: you need Node.js 18+ (LTS recommended) and an Optimizely DXP project with API credentials.

{
  "mcpServers": {
    "optimizely-dxp": {
      "command": "npx",
      "args": ["-y", "@jaxon-digital/optimizely-dxp-mcp"],
      "env": {
        "OPTIMIZELY_PROJECT_ID": "your-project-id",
        "OPTIMIZELY_PROJECT_KEY": "your-key",
        "OPTIMIZELY_PROJECT_SECRET": "your-secret"
      }
    }
  }
}

Additional setup notes

For HTTP-based automation platforms, you can run the MCP as an HTTP server with a JSON-RPC 2.0 API. The server exposes the /mcp endpoint for tool calls and /health for readiness checks. If you run via Docker, pass the required environment variables for your Optimizely project and enable HTTP transport.

If you prefer stdio (local development or Claude Desktop), you start the MCP by configuring a stdio-based MCP server in your client configuration. Claude will start and stop the MCP as needed, without manual intervention.

Configuration & environment variables

The MCP supports multiple environment variables to tailor deployment contexts, transport mode, and Redis caching. Common variables include project credentials and transport settings. Ensure you provide the correct IDs, keys, and secrets for your Optimizely DXP project.

If you enable HTTP mode, set the transport mode and port, for example DXP_MCP_MODE=http and DXP_MCP_PORT=3001. For Redis caching, you can enable Redis with REDIS_HOST, REDIS_PORT, REDIS_PASSWORD, and REDIS_ENABLED.

Troubleshooting & notes

Do not run the MCP server as a traditional background process. The MCP is started and stopped automatically by the client (Claude or your automation platform) and is designed to run on demand. If you encounter issues, verify credentials, ensure the DXP API access is enabled, and test the connection with the test tool.

Available tools

test_connection

Validate setup and show capabilities

check_permissions

Detailed environment access breakdown

verify_access

Confirm specific environment access

health_check

System status with structured health data

list_deployments

Show deployment history with filters

start_deployment

Initiate code deployment

monitor_deployment

Real-time progress with auto-refresh

complete_deployment

Finish verification state

reset_deployment

Rollback if needed

get_deployment_status

Current status with wait-then-check support

copy_content

Sync content between environments

list_content_copies

Show content copy history

export_database

Interactive export workflow with smart monitoring

check_export_status

Progress tracking with auto-download flag

download_database_export

Get export file with background progress

list_recent_exports

Export history and monitoring

analyze_logs_streaming

Stream and analyze in-memory for faster log analysis

compare_logs

Side-by-side log comparison for deployment decisions

download_logs

Download with manifest tracking

list_log_containers

Show available log containers

discover_logs

Find logs by date range and type

check_download_status

Progress tracking for active downloads

list_storage_containers

Show blob containers with structured data

download_blobs

Incremental blob downloads with parallelism

generate_storage_sas_link

Create temporary blob access URLs

list_download_history

Show completed downloads with manifests

list_projects

Show all configured projects

switch_project

Change active project context

current_project

Display active project info

get_ai_guidance

Context-aware best practices

get_version

Version info with update checks

get_download_paths

Show download configuration

set_download_path

Configure download paths by type

list_active_downloads

Progress for all background downloads

cancel_download

Stop background download

get_rate_limit_status

Show API quota and limits

get_cache_status

Redis cache statistics (if enabled)

monitor_project_upgrades

Track DXP CMS version updates

enable_http_logs

Configure HTTP log streaming

disable_http_logs

Disable HTTP log streaming

get_tool_availability

Show tool availability in current context

subscribe_deployment_events

NEW: MCP Resources for real-time updates