Testkube AI Configuration Quick Start

Cloud Control Plane Users

For users of the Testkube Cloud Control Plane, AI features are enabled as follows:

• AI Agents are always enabled for all environments in your organization.
• AI Assistant is disabled by default, enabled by one of the following methods (Admin or Owner only):
  1. When you first access the AI Assistant, an initial prompt is displayed with action buttons that allow you to either enable the feature or keep it disabled.
  2. Navigate to Organization Settings → Product Features tab and toggle AI Assistant on.

info

The default LLM used by AI functionality in Testkube Cloud is OpenAI's GPT-5.2-Codex.

You can add your own models (OpenAI, Azure OpenAI, or any OpenAI-compatible service) to make additional models available — see Configuring AI Models.

On-Prem Users

For On-Prem installations, you need to configure the AI service manually as described below - AND enable the AI Assistant in the Dashboard as described above.

AI Service Helm Configuration

Prerequisites:

• OpenAI API key (or compatible LLM API)
• PostgreSQL database (for conversation persistence) - this can live alongside your main Testkube PostgreSQL database.

testkube-ai-service:
  enabled: true
  postgres:
    # Database name (defaults to 'ai' if empty)
    database: "testkube_ai"
    secretRef:
      name: my-postgres-credentials     # Secret with DB credentials
      endpointKey: host                 # Key containing DB host
      usernameKey: username             # Key containing DB username
      passwordKey: password             # Key containing DB password
  inference:
    defaults:
      secretRef: my-openai-secret
      secretRefKey: api-key
    agent:
      - model: gpt-4o
        default: true
    tasks:
      model: gpt-4o-mini
    embeddings:
      model: text-embedding-3-small

testkube-cloud-api:
  aiApiSubdomain: ai
  restApiSubdomain: api

testkube-cloud-ui:
  ai:
    enabled: true

---

Optional: Custom LLM Endpoint

For Azure OpenAI, self-hosted models, or other OpenAI-compatible APIs:

testkube-ai-service:
  inference:
    defaults:
      extraHeaders:
        api-key: "your-azure-key"
      queryParams:
        api-version: "2024-02-01"
    agent:
      - model: gpt-4o
        default: true
        url: "https://my-company.openai.azure.com/openai/deployments/gpt-4o"
    tasks:
      model: gpt-4o-mini
      url: "https://my-company.openai.azure.com/openai/deployments/gpt-4o-mini"
    # Omit embeddings if your provider doesn't support /v1/embeddings

---

Default Values Reference

Setting	Default	Required for AI
testkube-ai-service.enabled	false	Must be true
testkube-cloud-ui.ai.enabled	false	Must be true
testkube-ai-service.postgres.*	empty	PostgreSQL required
testkube-ai-service.inference.defaults.secretRef	empty	API key required
postgresql.enabled	false	Only if using bundled DB

Next Steps

• AI Assistant — Start using the AI Assistant in the Dashboard
• AI Agents — Learn about default and custom AI Agents
• AI Agent Triggers — Automate agent execution on workflow events or schedules
• Configuring AI Models — Add custom models
• AI Configuration Reference — Advanced configuration, security, and compliance