Configuration Manuals

Configure AI/Run CodeMie to match your organizational requirements and workflows.

CodeMie

Data Sources Configuration

Configure data source loaders for code files, Confluence, Jira, and file uploads with custom chunking strategies and processing options.

→

Customer Feature Configuration

Show or hide UI components and configure custom URLs for the video portal, user guide, admin actions, and help resources.

→

Dynamic Customer Configuration

Change the banner, chat disclaimer, and other declared settings from the Administration page without a redeploy.

→

Code Executor

Configure sandbox mode, resource limits, and security policies for the isolated Kubernetes pods that run user-submitted Python code.

→

Scaling & Availability

Tune the Horizontal Pod Autoscaler and Pod Disruption Budget to control CodeMie API pod counts and availability during disruptions.

→

AI Models Integration

Configure LLM and embedding models from Azure OpenAI, AWS Bedrock, and Google Vertex AI with native or LiteLLM proxy integration.

→

API Configuration

Configure CodeMie API settings including application parameters, AI providers, database connections, security, storage, and observability.

→

Platform Administration

Create and manage projects, and access platform-wide settings as a global admin.

→

Project Budget Management

Enable and configure project-level budgets so admins can track and limit AI usage spend per project.

→

Access Control

Initial Realm Setup

Complete the one-time Keycloak setup required to store custom attributes before creating any users.

→

Platform-managed Mode Configuration

Enable Platform-managed mode for fresh installations or when switching from Keycloak-managed mode.

→

User Provisioning

Choose a method to create user accounts — manual creation, Keycloak Assistant, or Entra ID federation.

→

User Authorization

Grant users platform and project permissions by assigning roles, attributes, and guest mappings.

→

Extensions

Configure optional extensions to enhance AI/Run CodeMie capabilities.

✨ Assistants Evaluation

Data Volume Maintenance

Run ClickHouse maintenance queries for Langfuse — analyze disk usage, clean up data, and monitor TTL settings.

→

✨ LiteLLM Proxy

Budget Configuration

Configure budgets, rate limits, and spending controls in LiteLLM Proxy for cost management and usage tracking.

→

Model Configuration

Add and configure AI models in LiteLLM Proxy's config.yaml to route requests to multiple LLM providers through a unified gateway.

→

Spend Logs Retention

Manage the retention period for LiteLLM spend logs that track API calls, token usage, and cost per request.

→

Observability

Logs Retention

Manage retention for infrastructure logs aggregated in Elasticsearch to prevent unbounded disk usage.

→

Metrics Index Rotation

Configure quarterly rotation for the Elasticsearch index that stores CodeMie's user and usage metrics.

→

Before You Begin

Configuration Prerequisites

• Ensure AI/Run CodeMie is deployed and accessible
• Have administrative access to Keycloak and the platform
• Gather any credentials needed for external integrations (identity providers, data sources)

Getting Help

Each configuration guide provides detailed instructions and troubleshooting steps. For additional assistance, consult the FAQ or contact your support team.