Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/openagen/zeroclaw/llms.txt

Use this file to discover all available pages before exploring further.

ZeroClaw can manage itself as a background service using your operating system’s init system. On Linux distributions that run systemd, the service installs at the user level — no sudo required. On Alpine Linux (and other OpenRC-based distributions), the service installs system-wide and requires root. ZeroClaw auto-detects which init system is present and uses the correct backend automatically.
Init systemScopeConfig pathRequires sudo
systemd (default on Linux)User-level~/.zeroclaw/config.tomlNo
OpenRC (Alpine Linux)System-wide/etc/zeroclaw/config.tomlYes
Systemd is the default init system on most Linux distributions (Debian, Ubuntu, Fedora, Arch, and others). ZeroClaw installs as a user-level systemd unit, which means it runs under your own account with no elevated privileges.
1

Install the service unit

zeroclaw service install
This registers a zeroclaw.service unit file under ~/.config/systemd/user/ and enables it for your session.
2

Start the service

zeroclaw service start
3

Check service status

zeroclaw service status
4

Restart after config changes

zeroclaw service restart
5

Stop the service

zeroclaw service stop
6

Remove the service unit

zeroclaw service uninstall

Viewing logs on systemd

Systemd routes service output to the journal. Use journalctl to follow logs in real time:
journalctl --user -u zeroclaw.service -f
To view the last 100 lines without following:
journalctl --user -u zeroclaw.service -n 100

Config paths

The config file location differs between init systems. Both files use the same TOML schema; the path is determined by which init system is active.
api_key = "sk-..."
default_provider = "openrouter"
default_model = "anthropic/claude-sonnet-4-20250514"

[gateway]
port = 42617
host = "127.0.0.1"

Incident response

If the service becomes unhealthy, follow this sequence to restore it cleanly:
# 1. Snapshot current state
zeroclaw status
zeroclaw doctor
zeroclaw channel doctor

# 2. Check service status
zeroclaw service status

# 3. Restart cleanly
zeroclaw service stop
zeroclaw service start

# 4. Verify recovery
zeroclaw status
zeroclaw channel doctor
If channels continue to fail after a restart, verify allowlists and credentials in your config file.

Next steps

Running the daemon

Foreground daemon, gateway, agent modes, and cron scheduling

Network configuration

Tunnel providers, public bind options, and reverse proxy setup

Docker

Container-based deployment and sandboxed shell execution

Operations runbook

Health checks, log inspection, and rollback procedures

Build docs developers (and LLMs) love