1. Services Overview
2. Installation
   • Install with Docker Compose
     • Install Docker
     • Install Docker Compose
     • Download Docker Compose Files
     • Set Up Container Variables
     • Run GenAI Service Containers
     • Configure GenAI Functionality in Creatio
     • Notes
   • Install via Kubernetes
     • Deployment Overview
     • Prerequisites
     • Download Helm Files
     • Step-by-Step Configuration Guide
     • Deploy with Helm
     • Verify Deployment Status
     • Configure GenAI in Creatio

This deployment consists of several core services managed via Docker Compose. Each service plays a key role in the functioning of the GenAI system.

A web service responsible for handling GenAI requests such as chat completions.
It acts as the main REST API endpoint, processing input text from clients (e.g., Creatio) and returning responses generated by an underlying language model.

A service that provides an abstraction layer for Large Language Model (LLM) APIs.

Purpose:

• Enables easy integration with multiple LLM providers such as OpenAI and Azure OpenAI.
• Offers a unified API interface for making chat, completion, and embedding requests.
• Simplifies model management, routing, and provider switching with minimal configuration changes.

A PostgreSQL database used for storing:

• Operational statistics (e.g., request logs)

Install Docker on a physical or virtual machine running Linux OS to deploy the GenAI components.

Refer to the official Docker installation documentation for Linux (Debian-based distributions):
https://docs.docker.com/install/linux/docker-ce/debian/

Verify Docker installation by running the following command:

docker --version

1. Install Docker Compose on your Linux machine.

Refer to the official Docker Compose installation guide:
https://docs.docker.com/compose/install/

Verify Docker Compose installation by running:

docker-compose --version

1. Download and unpack the archive with the setup files from the repository:
   GenAI Docker Compose Setup - GitHub latest release

2. Navigate to the docker-compose folder

Configure the GenAI service containers by setting environment variables in the .env file located in the deployment folder. Edit this file to provide the required parameters.

Provide parameters depending on the Large Language Model (LLM) service you are using:

Set the default models used by GenAI (put model names of the provider that has to be used by default. You can take them from corresponding "openai" or "azure" sections):

Run this command, replacing your-username and your-password with your actual credentials:

docker login registry.creatio.com -u your-username -p your-password

1. Open a terminal and navigate to the docker-compose folder.

2. Run the following command to start the GenAI service containers in detached mode:

docker-compose up -d

1. Log into Creatio and go to System Settings.
2. Locate the setting: Account enrichment service url (code: AccountEnrichmentServiceUrl).
3. Set its value using the following format:

http://[your_server_ip_address]:5006

Replace [your_server_ip_address] with the actual IP address or hostname of the server where the GenAI Docker containers are running.

• This deployment uses Docker Compose for orchestration.
• Ensure that the firewall allows inbound traffic on port 5006 for Creatio to communicate with the GenAI service.
• Keep the .env file secure, as it contains sensitive API keys.

The Kubernetes deployment process consists of the following steps:

1. Prerequisites - Verify system requirements and tools (~5 minutes)
2. Download Files - Get the Helm chart files (~2 minutes)
3. Configuration - Configure your values.onsite.yaml file (~10-15 minutes)
4. Deploy - Install the application with Helm (~3-5 minutes)
5. Verify Deployment - Check deployment status and test API endpoint (~2 minutes)
6. Configure Creatio - Set up and test the connection in Creatio (~3 minutes)

Ensure the following components are set up:

• Kubernetes cluster (minimum version 1.19+)
• Helm package manager (minimum version 3.0+)
• kubectl configured to access your cluster
• Valid credentials for accessing registry.creatio.com

Run these commands to verify your setup:

Check Kubernetes cluster access:

kubectl cluster-info
kubectl get nodes

Check Helm version:

helm version

Check Docker registry access:

docker login registry.creatio.com -u your-username -p your-password

Expected output: All commands should complete successfully without errors.

• Kubernetes Setup Guide
• Helm Installation Guide
• kubectl Installation Guide

1. Download and extract the setup files from the repository:

   GenAI Kubernetes Setup - GitHub latest release

2. Navigate to the helm/genai directory within the extracted archive.

Follow these steps to configure your values.onsite.yaml file for deployment:

The configuration process involves setting up the following components in order:

1. Docker Registry Credentials (Required) - Access to pull container images
2. LLM Provider Configuration (Required) - OpenAI, Azure, or both
3. Service Access Configuration (Optional) - NodePort or Ingress setup
4. PostgreSQL Storage (Optional) - Enable persistent storage if needed

Important: You can configure parameters for either OpenAI, Azure, or both. If you are not using one of these providers, the corresponding section should be either commented out or removed entirely.

First, configure access to the Docker registry to pull container images:

dockerRegistry:
  username: <your-username>   
  password: <your-password>   
  email: <your-email>         

Replace with your actual credentials for accessing registry.creatio.com.

Choose one of the following options based on your LLM provider. Note: These are example configurations - you need to replace the placeholder values with your actual parameters.

Example configuration (replace with your actual values):

appConfig:
  genAI:
    llmProviders:
      models:
        openai:
          - name: gpt-4o
            model: gpt-4o
            api_key: sk-your-openai-api-key  # Replace with your actual OpenAI API key
          # - name: text-embedding-3-small  # Optional embedding model (disabled by default)
          #   model: text-embedding-3-small
          #   api_key: sk-your-openai-api-key  # Replace with your actual OpenAI API key
      defaultModel: gpt-4o
      # embeddingsModel: text-embedding-3-small # Optional - uncomment if using embeddings

Example configuration (replace with your actual values):

appConfig:
  genAI:
    llmProviders:
      models:
        azure:
          - name: azure-gpt-4o
            model: gpt-4o-2024-11-20  # Replace with your actual Azure model deployment
            resource_name: your-azure-resource  # Replace with your actual Azure resource name
            api_key: your-azure-api-key  # Replace with your actual Azure API key
          # - name: azure-text-embedding  # Optional embedding model (disabled by default)
          #   model: text-embedding-3-small  # Replace with your actual Azure embedding model
          #   resource_name: your-azure-resource  # Replace with your actual Azure resource name
          #   api_key: your-azure-api-key  # Replace with your actual Azure API key
      defaultModel: azure-gpt-4o
      # embeddingsModel: azure-text-embedding # Optional - uncomment if using embeddings

Parameter descriptions:

Important: The defaultModel and embeddingsModel values must exactly match the name field from the model definitions above.

By default, the service is accessible via NodePort on port 30082. Choose one of the following options:

No additional configuration needed. The service will be accessible on port 30082.

To change the port:

service:
  nodePort: 30082  # Change to your preferred port

To enable Ingress instead of NodePort:

ingress:
  enabled: true
  hosts:
    - genai.example.com  # Replace with your DNS name

By default, PostgreSQL runs without persistent storage enabled. This is normal behavior since the database only stores non-critical information such as request counts and token usage statistics.

If persistent storage is needed, add this to your configuration:

postgresql:
  primary:
    persistence:
      enabled: true

From the directory containing the values.onsite.yaml file, run:

helm upgrade --install genai . -f values.onsite.yaml

Before configuring Creatio, verify that your Kubernetes deployment is running successfully:

# Check if all pods are running (exclude only the flyway)
kubectl get pods

# Check service status
kubectl get services

Expected output: All pods should be in Running (exclude only the flyway) status and the service should show the NodePort (default: 30082).

Choose the appropriate URL format based on your configuration:

http://<kubernetes_node_ip>:30082

To find your Kubernetes node IP:

kubectl get nodes -o wide

http://genai.example.com

(Use the hostname you configured in your ingress settings)

For NodePort:

curl -X GET http://192.168.1.100:30082/readiness  # Replace 192.168.1.100 with your actual node IP

For Ingress:

curl -X GET http://genai.yourdomain.com/readiness

Expected response: HTTP 200 with health status information.

1. Log into Creatio as an administrator
2. Navigate to System Settings (System Designer → System Settings)
3. Search for code: GenAIServiceUrl
4. Set the value using the URL determined in the previous section
5. Save the configuration

Example configurations:

• NodePort: http://192.168.1.100:30082 # Replace 192.168.1.100 with your actual node IP
• Ingress: http://genai.yourdomain.com # Replace with your actual domain

1. Open Creatio AI Chat in your Creatio interface
2. Send a test message: "Hi, how are you?"
3. Verify response: You should receive a response from the AI assistant