Skip to main content
New to Testkube? Unleash the power of cloud native testing in Kubernetes with Testkube. Get Started >

Testkube AI Assistant On-Prem Installation

Testkube AI Assistant supports any LLM service that implements the OpenAI API specification, giving you full flexibility in where and how you run the model.

Configurations​

tip

Ensure that traffic to both your LLM endpoint and the Testkube AI Service API is allowed by your firewalls.

Setting Up API Access​

First, create a Kubernetes secret that holds the API key in the same namespace where the Testkube control-plane is installed:

kubectl -n <namespace> create secret generic <secret name> \
  --from-literal=LLM_API_KEY=<api key>

Configuring the Helm Chart​

In the testkube-enterprise chart you configure two components:

  1. testkube-ai-service – the backend that communicates with the LLM.
  2. testkube-cloud-ui – the web UI that surfaces AI Assistant features.
# Activate and wire the AI service
testkube-ai-service:
  enabled: true
  llmApi:
    # Optional URL for custom LLM API services
    # (defaults to OpenAI API when omitted)
    url: ""
    # K8s secret that stores your LLM_API_KEY
    secretRef: "<secret name>"

# Enable AI Assistant features in the UI and tell it
# where the ai-service is reachable
testkube-cloud-ui:
  ai:
    # -- Enable Testkube AI features
    enabled: true
    # -- Testkube AI service API URI
    aiServiceApiUri: "https://ai.<your-domain>"

Using OpenAI​

If you are connecting directly to OpenAI, leave llmApi.url empty and only reference the secret:

testkube-ai-service:
  enabled: true
  llmApi:
    secretRef: "<secret name>"

Using Self-Hosted LLMs or Third-Party Proxies​

For self-hosted models (e.g., vLLM, LiteLLM, OpenLLM) or commercial gateways that expose the OpenAI API:

testkube-ai-service:
  enabled: true
  llmApi:
    url: "http://your-llm-service:8000/v1"
    secretRef: "<secret name>"

Trying AI Assistant with the Testkube Hosted LLM Proxy (Trials & Demos)​

If you just want to explore AI Assistant without operating your own model, you can use our hosted LLM proxy at llm.testkube.io.
Request an evaluation license from testkube.io and point the AI service to the proxy:

testkube-ai-service:
  enabled: true
  llmApi:
    url: "https://llm.testkube.io" # hosted proxy
    # No secretRef needed – authentication is handled
    # via the Testkube license key you already supplied
warning

The hosted proxy is intended only for trials, demos, and onboarding. It has usage limits and is not recommended for production workloads.

Operating Behind an HTTP Proxy​

If egress traffic is routed through a corporate proxy, inject the proxy variables into the AI service pod:

testkube-ai-service:
  enabled: true
  llmApi:
    secretRef: "<secret name>"
  extraEnvVars:
    - name: HTTP_PROXY
      value: "http://proxy.domain:8080"
    - name: HTTPS_PROXY
      value: "https://proxy.domain:8043"
    - name: NO_PROXY
      value: ""