ClawFleet
ClawFleet is an open-source tool that runs a fleet of AI agents (OpenClaw and Hermes) in isolated Docker containers on your own Mac or Linux machine, managed from a single browser dashboard. It solves the problem of managing multiple AI agents with separate identities and skills without cloud costs, using your existing LLM subscriptions. It is for developers who want to deploy and manage their own AI workforce locally.
🎬 ClawFleet - Install in 30 Seconds · Desmond Wei
✨ Key features
- Unified dashboard for OpenClaw and Hermes runtimes
- Each agent runs in its own Docker container with isolation
- Supports any LLM provider: OpenAI, Anthropic, Google, DeepSeek, or ChatGPT subscription
- clawfleet shell drops into any instance's terminal
- Version pinning to lock tested runtime versions
- Character system and skill management with 52 built-in and 13,000+ community skills
🎯 Use cases
- Run a team of AI agents with distinct personas collaborating in Discord
- Deploy a personal learning agent (Hermes) that remembers and improves over time
- Manage multiple AI assistants for different channels like Slack and Telegram
- Clone a trained agent's personality and memory into new instances
- Watch and steer agents via browser-based noVNC desktop
📦 Installation
🧰 Requirements: macOS or Linux; Docker Desktop recommended on Mac, otherwise Colima auto-installed; an LLM API key or ChatGPT subscription.
curl -fsSL https://clawfleet.io/install.sh | sh
The install command will:
- Install Docker if needed (Colima on macOS, Docker Engine on Linux)
- Download and install the
clawfleetCLI - Pull the pre-built sandbox image (~1.4 GB)
- Start the Dashboard as a background daemon
- Open http://localhost:8080 in your browser
🚀 Usage
clawfleet create 1clawfleet listclawfleet configure <name>clawfleet shell <name>⚠️ Good to know
OpenClaw has more features surfaced in the Dashboard than Hermes; Hermes support is newer with limited Dashboard scope, and bringing both to full parity is an active direction.
❓ FAQ
What are the system requirements?
You need macOS or Linux. On Mac, Docker Desktop is recommended, but ClawFleet can auto-install Colima as an alternative.
How much memory does each agent use?
Idle memory is about 700 MB per OpenClaw instance and 140 MB per Hermes instance on an M4 MacBook Air, but OpenClaw can rise ~3x when actively browsing.
Can I use my ChatGPT subscription?
Yes, ClawFleet supports using your ChatGPT subscription as a model provider for OpenClaw instances.
How do I restrict the dashboard to localhost on Linux?
Stop the dashboard and restart with clawfleet dashboard start --host 127.0.0.1, then use an SSH tunnel to access it remotely.
📊 Repository
🤖 Overview, features, install steps and FAQ were generated from the project's README on Sep 4, 2026. Always check the original source before running commands.