33 MCP tools that let AI coding agents create, manage, and work inside dev containers, keeping your machine clean and your environment reproducible and deploy-friendly. Use local Docker, GitHub Codespaces, Azure, AWS, Google Cloud or Kubernetes seamlessly!
curl -fsSL https://github.com/aniongithub/devcontainer-mcp/releases/latest/download/install.sh | bash
AI agents need to run code somewhere. Today that's your machine. It shouldn't be.
When AI agents write code, they run it on your machine:
Give the agent a container. It builds there:
The agent orchestrates everything through MCP tools. You just approve.
Let your agent work beyond the confines of your hardware, or within it.
devcontainer CLI
Fast, simple, no cloud account needed. Perfect for local development with the official devcontainer spec.
Multi-provider
Docker, Kubernetes, AWS, Azure, GCP — one tool. DevPod handles provisioning, we handle orchestration.
Cloud-hosted
Full VS Code environment in the cloud. Prebuilds, fast startup, managed by GitHub. Just bring an auth handle.
The agent manages credentials without ever seeing a token.
auth_status("github")
Lists all keyring accounts with scopes
"You have 2 GitHub accounts. Which one?"
codespaces_create(auth: "github-you", ...)
MCP server resolves the handle on every call
Everything an agent needs to manage dev containers end-to-end.