Hermes · In Docker
Run Hermes as an always-on container that survives reboots and upgrades cleanly.
The whole model is one idea: the image is stateless, and one data volume holds everything. Your config, keys, sessions, skills, and memories live in ~/.hermes on the host, mounted into the container at /opt/data, so you upgrade by pulling a fresh image and your data is never touched. Run setup once (docker run ... setup), then run detached as a gateway with --restart unless-stopped and -p 8642:8642. The flags that matter: -v the data volume (required), -p 8642 (API + health), --restart unless-stopped, HERMES_DASHBOARD=1, and --memory/--cpus limits. Running several agents? Use one container per profile, each on its own data dir and port, never two gateways on one directory. Upgrade with docker pull + recreate.
Sources · What this video distills
1 docs page · every command below traces to one of themCommands shown · Copy and paste
each shows the source doc it came fromdocker run -it --rm -v ~/.hermes:/opt/data nousresearch/hermes-agent setupdocker run -d --name hermes --restart unless-stopped -v ~/.hermes:/opt/data -p 8642:8642 nousresearch/hermes-agent gateway rundocker pull nousresearch/hermes-agent:latest · docker rm -f hermes · (re-run)