browse it as a searchable site: https://notwitcheer.github.io/hermes-recipes/
working configs and scripts from my Hermes Agent walkthroughs. everything here ran on real hardware before it landed: a Mac Mini M4 that runs Hermes Agent 24/7, and an RTX 5090 Linux box for local inference.
Hermes Agent is Nous Research's open-source agent: autonomous skill creation, persistent memory with cross-session recall, cron jobs, mcp support, 20+ messaging gateways. docs: https://hermes-agent.nousresearch.com
recipes/has one folder per walkthrough: the exact configs and scripts I used, plus a readme with the commands and what they producedCHANGELOG.mdtracks what shipped in Hermes Agent and how to use it, in plain words- if a step has not been tested on real hardware, it does not ship here
live:
- 24/7 Hermes Agent on a Mac Mini M4: crons, approvals, and a gateway that stays up: the launchd gotcha and watchdog, the
[SILENT]cron pattern, a git-synced workspace, and the human-in-the-loop approval rule - Hermes Agent on a cheap VPS: zero to a Telegram gateway that survives reboot: the secure-box hardening script, the install one-liner, driving it interactively over tmux, and the systemd setup (with the #42065 install gotcha) that brings it back after a reboot
- lock down a self-hosted Hermes Agent: what it exposes and how to keep it shut: the gateway opens no inbound port, the dashboard refuses a public bind without auth, the API server is off by default, plus secret hygiene, backup/restore, and least-privilege scoping, with an exposure-audit script
coming next:
- extending Hermes Agent with a custom mcp server (TypeScript)
- persistent memory and fts5 recall in practice
watch the repo for new recipes. found a problem with one? open an issue, I fix what I ship.