Canonical guide · Source docs/getting-started.md
Getting started
Relmio 0.14.0 installs OAuth-backed AI connections. Pick the provider, location, and client that match your setup. SuperGrok is available for local apps and existing n8n deployments on the same computer or a VPS. It uses its own official device sign-in and never requires or reads ChatGPT credentials.
| Need | Choose | Credential |
|---|---|---|
| SuperGrok for a local backend | Grok Build adapter | Official subscription OAuth and a local bearer |
| SuperGrok for existing local or VPS n8n | Private SuperGrok companion | Official subscription OAuth and a local bearer |
| A trusted native Codex client | Codex App Server | ChatGPT sign-in and a local capability |
| A small local backend | Codex Chat Adapter | ChatGPT sign-in and a local bearer |
| A bridge for local Docker n8n or a VPS | n8n OAuth sidecar | A local ChatGPT sign-in file |
| n8n AI Assistant tools | Code Sandbox, with optional SearXNG | A generated sandbox key and a model credential entered in n8n |
ChatGPT sign-in is never converted into an OpenAI Platform API key. The Codex
routes are experimental and are not general /v1 services. The n8n OAuth
sidecar is unofficial, private, and policy-uncertain.
Install
On macOS, Linux, or WSL with Node.js 24 or newer:
npx --yes --ignore-scripts relmio@latest
The command opens the ChatGPT on my server route in a foreground browser
wizard without creating persistent Relmio state. It works when there is no
.relmio directory and no local n8n stack. Use relmio local only when you
explicitly want the setup route for this computer. Relmio checks prerequisites
in the browser, shows the plan, and asks before it writes files or starts
Docker. Local endpoints use 127.0.0.1. The n8n bridge and Assistant tools use
one selected Docker network and publish no host port. SearXNG is off by default.
On Git Bash, use the hosted launcher. It downloads a checksum-verified temporary
Node.js runtime and uses Git for Windows' bundled winpty bridge so the wizard
keeps its interactive terminal:
curl -fsSL https://relmio.jpfusin.tech/install.sh | sh
Direct NPX on Git Bash 2.38.1 still needs MSYS=enable_pcon for that one
process. Native PowerShell and Command Prompt installers are also available.
Keep the dashboard available
Install a persistent command with Node.js 24 or newer, then use its explicit lifecycle commands:
npm install --global --ignore-scripts relmio@latest
relmio start
relmio status
relmio open
relmio stop
relmio start runs the owner-scoped dashboard in the background. relmio status verifies only that process without printing its private session value.
relmio open starts it when needed and opens its private page. relmio stop
stops only that process; it does not stop or restart n8n, ngrok, endpoints,
bridges, Assistant companions, or unrelated containers.
The hosted curl, PowerShell, and Command Prompt launchers can use a verified
temporary runtime. Git Bash always uses that path so its native Node child can
run through winpty. In those cases the wizard remains a foreground, one-shot
process and ends with that terminal session. The temporary runtime is removed,
so install Relmio persistently before relying on these lifecycle commands.
Choose a guide
- Local dashboard for launch commands, service states, available actions, and credential boundaries.
- Local endpoints for the gateway, Codex, the Chat Adapter, the n8n bridge, and local Assistant tools.
- New local n8n + ngrok if you do not already run n8n.
- AI Assistant companion for Assistant setup and its limits.
- VPS and n8n for the remote sidecar route.
- SuperGrok on a VPS for the provider-specific VPS flow.
- Troubleshooting when setup stops.
- Security for account, network, and credential rules.