Skip to content
devopsellence docs

devopsellence docs

AI-operator-first deployments on vanilla VMs.
Quickstart

devopsellence helps AI coding and operations assistants deploy containerized apps without inventing production shell choreography.

Why care today: it gives Codex, Claude, or a human operator a narrow deployment contract for Dockerized apps on VMs. Plan a change, deploy it, verify status and HTTPS, inspect logs, manage secrets, and roll back with structured evidence instead of guesses.

The contract stays narrow: inspect, plan, apply desired state, observe reconciliation, and recover with ordinary tools when needed. Humans stay in the approval loop; the node agent stays deterministic.

Start here

Section titled “Start here”

Already have a Dockerized app:

Terminal window
curl -fsSL https://www.devopsellence.com/lfg.sh | bash
~/.local/bin/devopsellence skill install --global
cd my-app
codex "deploy with devopsellence solo"

devopsellence skill install installs the matching AI agent skill from the CLI itself.

  • Solo quickstart for the shortest path to one VM.
  • Ingress and TLS for hostnames, DNS checks, and HTTPS verification.
  • Basecamp Fizzy on Rails for a real Rails app example that maps a Kamal-style deployment to devopsellence solo.
  • Flue agents on Node.js for deploying an experimental Flue webhook agent server as an ordinary containerized service.
  • Sessy for Amazon SES for running an SES observability dashboard on a VM while your app sends through Amazon SES.
  • AI operator model for the product thesis: a CLI that gives AI operators structured feedback, safe boundaries, and facts they can compose with the user’s tools.
  • Runtime model for desired state, releases, services, nodes, and status.
  • CLI reference for the AI-operator-safe command surface.

Solo and shared use the same runtime model. Mode changes ownership, persistence, and transport; it should not change deployment semantics.