The ultimate harness for coding agents and beyond. All your agents. All your machines. One command center. Start with code, then follow your curiosity and build across disciplines: CAD, circuits, robots, games and music.
TypeScript 3d-modelingagent-frameworkagent-harnessai-agentscadclaude-codecodexcoding-agentsdeveloper-toolsesp32game-developmentmcp-servermulti-agentopen-hardwareopencodeopenharnesspcbroboticsterminaltmux
[Unit]
Description=Harness agent runtime
ConditionPathExists=!/etc/harness-live
[Service]
Type=exec
Environment=ADAPTER_UPDATE_DISABLE=true
Environment=HARNESS_OS=1
# `start --foreground` runs harnessd's master, as launchd does; here it keeps running the core alone, one
# process, as the memory and zram checks in os/tests measure it (self-update is off, so no update needs
# the master to judge it).
Environment=HARNESS_NO_MASTER=1
Environment=HARNESS_TUI_BIN=/usr/lib/harness/hn
Environment=PATH=%h/.local/bin:/usr/local/bin:/usr/bin
Environment=npm_config_prefix=%h/.local
Environment=BROWSER=hn-browser
ExecStart=/usr/bin/harness start --foreground
ExecStartPost=/usr/lib/harness-os/wait-runtime
TimeoutStartSec=150
Restart=on-failure
RestartSec=2
KillMode=process
# The core alone ends its own stop after 2.5 s (cli/src/core/stopDeadline.ts). An older CLI has no such
# bound, and systemd's default 90 s held an update for all of it (2026-10-13). Its main process only, as
# the CLI's own unit does: the agents in tmux outlive the kill.
TimeoutStopSec=10
[Install]
WantedBy=default.target