HolyClaude logo

HolyClaude

Free

AI coding workstation: Claude Code + web UI + 7 AI CLIs + headless browser + 50+ tools

FreeFree tier
Type
Open Source

About HolyClaude

HolyClaude is an AI coding workstation packaged as a Docker image, providing a pre-configured environment with a CloudCLI web UI, headless browser automation via Playwright, and multiple AI coding agents accessible through CLI. It supports AMD64 and ARM64 architectures and offers two image variants: full (8 AI CLIs) and slim (5 AI CLIs). The included AI CLIs are Claude Code, Gemini CLI, OpenAI Codex, Cursor Agent, TaskMaster AI, Junie, OpenCode, and Pi Coding Agent (full only). The container runs locally by default on port 3001 with mapped host folders for settings and workspace. Additional features include rootless Podman support, optional SSH/Mosh remote shell, reverse-proxy path configuration, account controls (logout, password change, token invalidation), CloudCLI plugins (Project Stats, Web Terminal), and notifications via Apprise. The project is open source under the MIT license.

Key Features

CloudCLI web UI with password/token control
Multiple AI CLI tools: Claude Code, Gemini CLI, OpenAI Codex, Cursor Agent, TaskMaster AI, Junie, OpenCode, Pi Coding Agent (full only)
Headless browser automation via Playwright with Debian Chromium
Two image variants: full (8 AI CLIs) and slim (5 AI CLIs)
Supports AMD64 and ARM64 platforms
Rootless Podman support with keep-id and SELinux :Z mounts
Optional remote shell via SSH and Mosh
Reverse-proxy path support (HOLYCLAUDE_BASE_PATH)
CloudCLI plugins: Project Stats and Web Terminal
Notifications via Apprise for Claude, Codex, Gemini, and CloudCLI Codex

Pros & Cons

Pros
  • Pre-packaged environment with multiple AI tools ready to use
  • Supports both AMD64 and ARM64 architectures
  • Option for slim image to reduce download size (2.55 GB compressed)
  • Built-in headless browser for web automation without extra setup
  • Can be run rootless with Podman for better security compatibility
  • Easy to update via docker compose pull
  • Configurable reverse-proxy path for hosting behind a URL prefix
  • Account controls (logout, password change, token invalidation) included in CloudCLI
Cons
  • Full image variant is large (4.38 GB compressed transfer size)
  • Single-user only; no multi-tenant isolation
  • CloudCLI self-update is patched out; requires pulling a new image
  • Requires Docker or Podman with enhanced privileges (SYS_ADMIN, SYS_PTRACE, seccomp unconfined) that reduce security isolation
  • Mounted files, credentials, added capabilities, and outbound provider access remain reachable within container limits

Best For

AI-assisted coding and developmentAutomated code generation and reviewBrowser automation for testing or scrapingLocal development environment with multiple AI agentsRemote coding workstation (SSH/Mosh)Running AI agents in a repeatable Docker workspace

FAQ

What AI CLIs are included in HolyClaude?
The full image includes 8 AI CLIs: Claude Code, Gemini CLI, OpenAI Codex, Cursor Agent, TaskMaster AI, Junie, OpenCode, and Pi Coding Agent. The slim image includes 5: Claude Code, Gemini CLI, OpenAI Codex, Cursor Agent, and TaskMaster AI.
How do I start HolyClaude?
Create a docker-compose.yaml file with the provided configuration, create data/claude and workspace directories, then run 'docker compose up -d'. Open http://localhost:3001 to access the CloudCLI web UI.
What is the difference between the full and slim images?
The full image (4.38 GB compressed) includes 8 AI CLIs, while the slim image (2.55 GB compressed) includes 5 AI CLIs. Both include CloudCLI, headless browser, and other tools.
How do I update HolyClaude?
Run 'docker compose pull' followed by 'docker compose up -d' to pull the latest reviewed image. CloudCLI self-update is patched out in favor of pulling new images.
Can I use HolyClaude with Podman?
Yes, rootless Podman is supported. Use the rootless instructions with keep-id and SELinux :Z mounts for host-owned workspaces.
Is HolyClaude free to use?
Yes, HolyClaude is open source under the MIT license and is free to use. You need your own credentials for the AI providers you choose to use.