Authentication:GHE username + personal access token (repo scope)
OpenShift CR:ProjectHelmChartRepository with basicAuthConfig
🔒 Internal access requires GHE credentials.
This site is SSO-gated. Use raw.github.ibm.com with your GHE username and a personal access token (repo scope) for both Helm CLI and OpenShift.
Kubernetes operator for deploying and managing IBM Workflow and Document Understanding (WDU) Services—enabling
intelligent document extraction, automated classification, and workflow orchestration for enterprise content.
Chart Version:26.1.1
App Version:26.0.1
Architecture:amd64
Key Features
Intelligent document understanding and field extraction
AI-powered document classification and routing
Automated workflow trigger based on document content
Deep integration with IBM Content Platform Engine
Secure credential management via Vault integration
Cloud-native operator pattern with namespace-scoped RBAC
Deploys Core MCP server and Reasoning service with Navigator AI Agent. Enables intelligent content processing,
agentic AI capabilities, and natural language interactions with enterprise content.
Chart Version:26.1.1
App Version:26.0.1
Architecture:amd64
Key Features
Core MCP (Model Context Protocol) server
Reasoning service for intelligent decision-making
Navigator AI Agent for agentic content interactions
Redaction MCP — AI-driven content redaction via MCP protocol
LegalHold MCP — legal hold lifecycle management via MCP protocol
Kubernetes operator for deploying and managing IBM Workflow and Document Understanding (WDU) Services—enabling
intelligent document extraction, automated classification, and workflow orchestration for enterprise content.
Chart Version:26.1.0
App Version:26.0.1
Architecture:amd64
Key Features
Intelligent document understanding and field extraction
AI-powered document classification and routing
Automated workflow trigger based on document content
Deep integration with IBM Content Platform Engine
Secure credential management via Vault integration
Cloud-native operator pattern with namespace-scoped RBAC
Architecture Overview
WDU (Workflow and Document Understanding) is a pipeline-based microservice stack that sits between the content repository and AI models. Each document traverses the pipeline stages shown below.
Deploys Core MCP server and Reasoning service with Navigator AI Agent. Enables intelligent content processing,
agentic AI capabilities, and natural language interactions with enterprise content.
Chart Version:26.1.0
App Version:26.0.1
Architecture:amd64
Key Features
Core MCP (Model Context Protocol) server
Reasoning service for intelligent decision-making
Navigator AI Agent for agentic content interactions
Redaction MCP — AI-driven content redaction via MCP protocol
LegalHold MCP — legal hold lifecycle management via MCP protocol
Kubernetes operator for deploying and managing IBM Redis CP—a high-availability Redis cache layer for IBM Content Cortex and IBM Cloud Pak for Data workloads.
Supports namespace-scoped deployment, credential secret management, persistent storage, and HAProxy configuration.
Manages PostgreSQL clusters embedded in IBM products. Provides automated provisioning, lifecycle management,
and backup/recovery for PostgreSQL instances used by IBM Content Cortex components.
Kubernetes operator for deploying and managing IBM Model Gateway—an LLM proxy that provides a unified,
protocol-normalized gateway layer for AI model access, intelligent routing, and lifecycle management
within IBM Content Cortex and Cloud Pak for Data environments.
Chart Version:13.0.6
App Version:13.0.6
Architecture:amd64, s390x, ppc64le
Key Features
LLM proxy with unified API surface (OpenAI-compatible)
Intelligent request routing across multiple model providers
Model lifecycle management and version pinning
Token usage tracking and rate-limit enforcement
Streaming and non-streaming response support
Multi-architecture support (amd64, s390x, ppc64le)
Architecture Overview
Model Gateway acts as an LLM proxy sitting between IBM Content Cortex services (AI Services, WDU) and the backing model providers. All LLM traffic is normalized to a single OpenAI-compatible API surface, decoupling service code from provider-specific SDKs.
Model Gateway — LLM Proxy Flow
🤖 AI Services Operator
📄 WDU Services Operator
▼ OpenAI-compatible REST
🔀 IBM Model Gateway (LLM Proxy)
Route by model ID / provider policy / availability
☁️ IBM watsonx.ai
🏠 On-premises LLM
🔌 3rd-party Provider
Key Design Properties
Protocol Normalization
Exposes a single OpenAI-compatible /v1/chat/completions endpoint regardless of the backing provider, eliminating SDK sprawl in consumer services.
Routing Policy Engine
Routes requests by model ID, namespace policy, latency SLA, and provider availability. Supports fallback chains for resilience.
Token Metering
Tracks prompt and completion token counts per request, namespace, and model. Feeds into IBM License Service for chargeback.
Credential Isolation
Provider API keys are stored as Kubernetes secrets, injected at the proxy layer. Consumer services never handle raw credentials.
Manages IBM License Service for tracking and reporting license usage across the entire cluster.
Required for compliance and license management of IBM Cloud Pak components.
Deploys Core MCP server and Reasoning service with Navigator AI Agent. Enables intelligent content processing,
agentic AI capabilities, and natural language interactions with enterprise content.
Chart Version:26.0.4
App Version:26.0.0
Release Date:2026-10-30
Architecture:amd64
Key Features
Core MCP (Model Context Protocol) server
Reasoning service for intelligent decision-making
Navigator AI Agent for agentic content interactions
Deploys Core MCP server and Reasoning service with Navigator AI Agent. Enables intelligent content processing,
agentic AI capabilities, and natural language interactions with enterprise content.
Chart Version:26.0.3
App Version:26.0.0
Release Date:2026-09-30
Architecture:amd64
Key Features
Core MCP (Model Context Protocol) server
Reasoning service for intelligent decision-making
Navigator AI Agent for agentic content interactions
Deploys Core MCP server and Reasoning service with Navigator AI Agent. Enables intelligent content processing,
agentic AI capabilities, and natural language interactions with enterprise content.
Chart Version:26.0.2
App Version:26.0.0
Release Date:2026-08
Architecture:amd64
Key Features
Core MCP (Model Context Protocol) server
Reasoning service for intelligent decision-making
Navigator AI Agent for agentic content interactions
Deploys Core MCP server and Reasoning service with Navigator AI Agent. Enables intelligent content processing,
agentic AI capabilities, and natural language interactions with enterprise content.
Chart Version:26.0.0
App Version:26.0.0
Release Date:2026-06-26
Architecture:amd64
Key Features
Core MCP (Model Context Protocol) server
Reasoning service for intelligent decision-making
Navigator AI Agent for agentic content interactions
Deploys Core MCP server and Reasoning service with Navigator AI Agent. Enables intelligent content processing,
agentic AI capabilities, and natural language interactions with enterprise content.
Chart Version:26.0.1
App Version:26.0.0
Release Date:2026-07-01
Architecture:amd64
Key Features
Core MCP (Model Context Protocol) server
Reasoning service for intelligent decision-making
Navigator AI Agent for agentic content interactions
Manages IBM License Service for tracking and reporting license usage across the entire cluster.
Required for compliance and license management of IBM Cloud Pak components.
git clone https://github.com/ibm-ecm/ibm-content-cortex-containers.git
cd ibm-content-cortex-containers/scripts
# Create and activate virtual environment
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
# Install dependencies
pip install -r requirements.txt
3
Deploy Operators
# Interactive deployment with guided configuration
python deploy_operator.py
# Follow prompts to:
# - Select operators to deploy
# - Configure namespaces
# - Set image pull secrets
# - Verify prerequisites
🔒 Dev mode — GITHUB_TOKEN required.
Running with --dev pulls pre-release chart artifacts from the internal GHE repository.
Export your GHE personal access token (repo scope) before running the command below.
🔒 IBM GHE credentials required.
The Pages URL is SSO-gated. OpenShift must pull chart archives from
raw.github.ibm.com using basic auth credentials.
See Step 2 below to create the credentials secret first, then apply this config in the same namespace:
🔒 IBM GHE credentials required.
The Pages URL is SSO-gated and cannot be used with Helm CLI.
Use the raw content URL with your GHE username and a personal access token (repo scope).