Kubeopencode
Run OpenCode Agent on Kubernetes
Last verified:
What is Kubeopencode?
KubeOpenCode is a Kubernetes-native AI agent platform designed for teams and enterprises to run AI coding agents on Kubernetes. Built on OpenCode, it serves as infrastructure that connects users to AI Code Agents through Kubernetes without building the AI engine itself. The platform allows teams to deploy persistent AI agents that run as Deployments with zero cold start, share context across tasks, and maintain session history that survives restarts.
Key features include a browser-based web terminal for accessing agents without CLI setup or SSH, auto suspend/resume (Standby) to save compute when agents are idle, CronTask for scheduled recurring task execution, Git Auto-Sync with hot-reload capability, concurrency and quota controls for task rate-limiting, and Skills from Git repos that are auto-injected as slash commands across all agents. The platform uses fully declarative CRDs that are GitOps-friendly and work with Helm, Kustomize, and ArgoCD.
KubeOpenCode is for development teams, DevOps engineers, and enterprises that want shared AI coding infrastructure rather than individual developer tools. It supports CI/CD integration through kubectl, Helm, GitHub webhooks, Slack bots, and any controller. The platform is ideal for organizations already using Kubernetes who want to scale AI coding agents across teams with proper isolation, credentials management, and lifecycle control.
The architecture is simple with no external dependencies—no PostgreSQL or Redis required, just Kubernetes with etcd for state. It uses a two-container pattern where an init container copies the OpenCode binary and a worker container runs the agent, all packaged in a single container image. The platform separates Task (what to do), Agent (how to execute), and Pod/OpenCode (where it runs).
Kubeopencode pricing
Pricing model: Freemium
KubeOpenCode itself is free and open-source under the Apache 2.0 license. There are no paid plans from the project. Users pay for their own infrastructure (Kubernetes cluster such as EKS, GKE, AKS, or local clusters with Kind/minikube) and any AI model API keys they choose to use from providers like Anthropic, OpenAI, Google Gemini, Amazon Bedrock, Azure, Google Vertex, xAI, Mistral, Groq, OpenRouter, or GitHub Copilot.
Kubeopencode pros
- Open-source and free under Apache 2.0 license
- Runs on any standard Kubernetes cluster v1.26+
- Supports 75+ AI providers including Anthropic, OpenAI, Gemini
- Zero cold start with persistent agent Deployments
- Browser-based web terminal no SSH or port-forwarding needed
- Auto suspend/resume saves compute when agents idle
- Session history persists across restarts
- Git Auto-Sync with hot-reload without restart
- CronTask for scheduled recurring task execution
- Fully Kubernetes-native declarative CRDs
- GitOps-friendly works with Helm Kustomize ArgoCD
- No external dependencies no PostgreSQL or Redis required
- Concurrency and quota controls with rate limiting
- Skills from Git repos shared across all agents
- Task API works with kubectl CI/CD webhooks Slack bots
- Single container image for controller and utilities
- No vendor lock-in on AI model or engine
Kubeopencode cons
- Currently in early alpha v0.1.x with unstable API
- API may change without backward compatibility
- Recommended only for development testing evaluation not production
- Users pay for their own Kubernetes infrastructure costs
- Users pay for their own AI model API keys
- Requires Kubernetes cluster knowledge to deploy
- Early stage project with limited community adoption
- Dependent on OpenCode engine capabilities and limitations
Frequently asked questions about Kubeopencode
What Kubernetes versions does KubeOpenCode support?
KubeOpenCode runs on any standard Kubernetes cluster version 1.26 or higher. You can use managed services like EKS, GKE, AKS, or a local cluster with Kind or minikube for development purposes.
Which AI models and providers are supported?
KubeOpenCode uses OpenCode as its AI engine, supporting every provider and model that OpenCode supports. This includes Anthropic, OpenAI, Google Gemini, Amazon Bedrock, Azure, Google Vertex, xAI, Mistral, Groq, OpenRouter, GitHub Copilot, and more. There are 75+ supported providers total, and you can also use any OpenAI-compatible endpoint.
Is KubeOpenCode free to use?
Yes, KubeOpenCode itself is free and open-source under the Apache 2.0 license. There are no paid plans or subscription fees from the project. However, you pay for your own Kubernetes infrastructure and any AI model API keys you choose to use.
What is the current development stage of KubeOpenCode?
KubeOpenCode is currently in early alpha (v0.1.x). The API may change without backward compatibility. The project recommends using it for development, testing, and evaluation while they work toward a stable release.
How do I access agents without using the CLI?
KubeOpenCode includes a built-in web terminal that lets you access any agent directly from your browser. This provides real-time interactive sessions without needing SSH or port-forwarding, making it accessible for team members who prefer not to use the command line.
What is the difference between Agent and Task resources?
Agent resources deploy persistent AI agents your team can interact with in real time through the web terminal, CLI, or by submitting Tasks. They have zero cold start and persist session history. Task resources run stable, repeatable AI tasks in ephemeral Pods, perfect for CI/CD pipelines, batch operations, and automated workflows using just kubectl apply.
How does Git Auto-Sync work?
Git Auto-Sync uses HotReload to sync Git contexts in-place without restarting the agent. The rollout policy triggers rolling updates with active task protection, ensuring ongoing tasks complete safely while the agent updates to the latest code context.
What external dependencies does KubeOpenCode require?
KubeOpenCode has no external dependencies. It does not require PostgreSQL, Redis, or any other external services. It uses only Kubernetes with etcd for state storage and Pods for execution, making it simple to deploy and maintain.
How do I get started with KubeOpenCode?
You can get started in minutes with Helm. The project provides Helm charts to deploy your first agent quickly. You can also view the GitHub repository for documentation and source code, and watch the demo video to see KubeOpenCode in action before deploying.
Can I use KubeOpenCode for CI/CD automation?
Yes, KubeOpenCode is designed for CI/CD integration. The Task API is a standard Kubernetes CRD triggerable from kubectl, Helm, CI/CD pipelines, GitHub webhooks, Slack bots, or any controller. CronTask enables scheduled automation for recurring work, and you can scale with Helm templates for batch operations.