Related ToolsClaude CodeCursor

HolyCode Packages OpenCode AI Agent with 30+ Dev Tools in One Docker Container

AI news: HolyCode Packages OpenCode AI Agent with 30+ Dev Tools in One Docker Container

What happens when you need your AI coding setup on a new machine? You reinstall Node, clone configs, set up Playwright, configure your AI provider keys, and lose an afternoon. HolyCode is a Docker container that eliminates that cycle by bundling the OpenCode AI coding agent with everything a developer might need, pre-installed and ready to go.

What's in the Box

The container ships with 30+ tools organized across categories:

  • Git and version control: git, lazygit, GitHub CLI, delta for readable diffs
  • Search and navigation: ripgrep, fd, fzf, bat (syntax-highlighted file viewer)
  • Browser automation: Chromium, Xvfb (virtual display server), Playwright
  • Runtimes: Node.js 22 LTS, Python 3
  • System tools: htop, tmux, curl, wget, jq, vim, ssh

The AI coding agent itself is OpenCode, which provides a web UI (accessible on port 4096) and supports multiple AI providers: Anthropic Claude, OpenAI, Google Gemini, Groq, AWS Bedrock, and Azure OpenAI. You pick your provider by setting an environment variable.

The Persistence Problem It Solves

The real selling point isn't the tool list. It's that all state - sessions, settings, MCP configurations (the protocol AI tools use to connect to external services), plugins, and tool history - lives in a bind mount outside the container. Rebuild the container, update the image, switch machines: your entire coding environment picks up where you left off.

For teams running multiple AI coding agents across different machines, or developers who frequently switch between workstations, this is genuinely useful. The UID/GID remapping means file permissions stay correct on bind mounts, which is one of those Docker headaches that's easy to get wrong.

Who This Is For

Developers who want a reproducible AI-assisted coding environment without managing the dependencies themselves. The project is MIT-licensed, supports Linux (amd64/arm64), macOS via Docker Desktop, and Windows through WSL2. A managed cloud version is listed as "coming soon" for those who'd rather skip Docker entirely.