Purpose
devops-mcp-webui is designed to turn the daily operations and maintenance of a Kubernetes cluster into natural language conversations. Users do not need to memorize complex kubectl or helm commands; they only need to ask questions in the Open WebUI chat interface, and the AI will automatically select the appropriate tool to execute and return formatted results. For example, ask "Which Pods are running in kube-system?" or "Scale all deployments to 5 replicas."
Main Capabilities
- Query Kubernetes resources through natural language, such as listing Pods, Services, Deployments, etc.
- Perform kubectl operations: kubectl_get (list resources), kubectl_apply (apply manifests), kubectl_describe (describe resources), etc.
- Perform helm operations, such as helm_install to install a Helm Chart.
- Support scenarios such as creating namespaces, viewing logs, and installing applications (such as Jenkins, ArgoCD).
- Compatible with Open WebUI's tool calling mechanism, and can display available tool icons in the chat input area.
- Support local Ollama models (such as llama3.2:latest) or connect to remote models such as OpenAI and Anthropic.
How to Use
- Use the one-click Docker Compose script provided by the project to start all components, including Open WebUI (http://localhost:3000）、MCP Bridge（http://localhost:9000）、Ollama API（http://localhost:11434）以及 Kind (Kubernetes in Docker) cluster.
- Open in the browser http://localhost:3000，首次访问创建管理员账户并登录。
- Go to Admin Panel → Tools, click "+ Add Tool Server", fill in the name "Kubernetes Tools", URL "http://mcpo:9000”、描述“Kubernetes management via kubectl and helm".
- Configure the model connection in Admin Panel → Connections: local Ollama is usually automatically detected (verify http://ollama:11434），也可添加远程提供商的 API key and endpoint.
- Return to the chat interface, click the tool icon (🔧) to confirm that the list of available tools appears (such as kubectl_get, kubectl_apply, kubectl_describe, helm_install, etc.).
- Enter natural language instructions directly in the chat, and the AI will automatically call the corresponding tool to execute and return the result.
Applicable Scenarios
- Developers and operations personnel who want to quickly view or manage a Kubernetes cluster through conversation.
- Learning or demonstrating Kubernetes operations on a local Kind cluster.
- Teams that need to integrate an AI chat interface with existing kubectl/helm workflows.
Limitations
- The project materials do not explain the supported production environment scale, security, or permission control details.
- Information about fees, pricing, platform compatibility, performance metrics, etc. is not yet provided.
- Kubernetes cluster access credentials (kubeconfig) must be prepared by the user; the project uses them through mounted volumes.
- The specific configuration and costs of remote model connections depend on the corresponding provider and are not expanded on in the materials.
