ClawPier

What ClawPier does

ClawPier runs every agent in a Docker container that starts with --network none, so network access is opt-in per bot. CPU and memory caps are set from the GUI, a one-click stop kills a misbehaving agent, and the interface covers chat, config, an interactive terminal, live logs, a file browser and resource monitoring without needing the CLI.

When creating a bot you choose the OpenClaw or Hermes runtime, and the dashboard shows the model, provider, platforms and settings for that agent type. Other features include auto-start, health checks with automatic restart, port-mapping presets, per-bot environment variables, and a ClawHub skill browser with 50+ bundled skills. It is built with Tauri v2, a Rust backend with a React frontend, and ships for macOS, Linux and Windows. The README says macOS builds are code-signed and notarized.

Key features

• Docker containers started with --network none by default
• Choice of OpenClaw or Hermes runtime per bot
• CPU and memory limits set from the GUI
• Interactive terminal, live logs, file browser and resource monitoring
• Health checks with automatic restart and optional auto-start
• ClawHub skill browser with one-click install

When to use it

• Running a Hermes agent in a container that cannot reach your host files or network
• Managing several OpenClaw and Hermes bots from one dashboard
• Stopping a misbehaving agent instantly without touching Docker flags

Who it is for: Users who run agent runtimes with broad system access and want a GUI-managed sandbox instead of the Docker CLI.

How it fits with Hermes Agent

Supports Hermes Agent alongside OpenClaw as a selectable runtime for sandboxed bots.

How to install ClawPier

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

brew tap SebastianElvis/clawpier
brew install --cask clawpier

Requirements: Docker installed and running, plus an OpenClaw or Hermes agent image, which ClawPier prompts you to pull on first launch

FAQ