spawnd · September 2026

Guides for the agent era

How the CLI coding agents actually work, how to reach the machines they run on, and what to do when either misbehaves — written to be useful whether or not you ever install anything of ours.

What these pages are for.

A coding agent is a program that takes a goal, works in a real shell on a real machine, and asks you for a yes now and then. Almost everything worth knowing about running one is unglamorous: how it authenticates, where its settings live, how to keep it running after the laptop closes, how to reach it from a phone, why a tunnel says “starting” forever. The vendors’ docs cover their own product on its happy path. These guides cover the seams between products, machines, and devices — the places where a search engine is the only support desk.

Every page here follows one rule: teach the thing you searched for first, completely, using tools you already have. Where that honest route runs out — a session that must survive your absence, a machine you would rather not expose, a phone that needs a real terminal — the page says so, and only then shows what spawnd does about it. If you stop reading before that turn, you still got what you came for.

How the pages are grouped.

Agent guides cover Claude Code and Codex in depth — plans and pricing, settings and commands, running them remotely and in teams — plus the open alternatives, OpenCode and Aider. Device guides cover SSH from an iPhone, iPad, or Android, and every method for reaching a Mac from somewhere else. Reference pages define the vocabulary the industry is settling on. Fix pages start from the exact error you pasted into the search box. Two tools do sums and lookups a page of prose can’t. Each page links its siblings, so the rack below is a map, not a menu.

Every guide, reference, definition, fix, and tool.

Claude Code

Plans, pricing, remote, teams, open source, settings, commands — the honest route first.

Read

Codex CLI

OpenAI’s terminal agent, end to end: install, sandbox, config, sessions, and where one machine runs out.

Read

Claude Code, remote

Remote Control, the web, and SSH + tmux — then the machines they leave out.

Read

Claude Code on Max

The 5x and 20x tiers, the two limit clocks, and how to make one plan stretch.

Read

Is Claude Code free?

Free to install, not free to run — every way to pay, side by side.

Read

Claude Code for teams

The Team and Enterprise plans, then the checked-in files and the machine problem.

Read

Claude Code open source?

It isn’t — here are the agents that are, and the open stack around it.

Read

Claude vs Claude Code

The assistant, the agent, and the machine that tells them apart.

Read

How to run Claude Code

Install, sign in, first task, the loop, and where it stops.

Read

Claude Code settings

Every settings.json, which one wins, and the keys worth setting.

Read

Claude Code commands

Slash commands, CLI flags, shortcuts, and your own.

Read

Not available in your country

What the message means, the fixes in order, and what no fix will do.

Read

Install Codex CLI

macOS, Linux, Windows native or WSL2 — sign-in, updating, and the step-one errors.

Read

Codex CLI sessions

Where every chat is saved, how to resume and export it, and what the files cannot keep alive.

Read

OpenCode vs Aider

An agent and a pair programmer, compared by a party that runs both.

Read

Is OpenCode free?

The program is; the model is the bill. Every way to pay it, including the free ones.

Read

Agentic orchestration

The definition worth keeping, the patterns behind it, and where the agents run.

Read

Cloud development environment

The definition, the history, the bill, and the self-hosted kind.

Read

Background agents

The definition, the products that use the term, and whose machine the agent is on.

Read

SSH from an iPhone

The client, the key, the reachability, the tmux — and the route with no client at all.

Read

SSH on an iPad

Clients, keys, the hardware keyboard, Stage Manager — and the honest meaning of a dev machine.

Read

SSH from Android

Termux from F-Droid, ConnectBot, Termius, the battery manager — and the route with nothing on the phone.

Read

Best SSH client for Windows

Built-in OpenSSH, PuTTY, Termius, MobaXterm, Tabby, Bitvise, WSL — and when a browser is the answer.

Read

Remote access to your Mac

Every method compared: what each opens, what each costs, and which work from a phone.

Read

Remote Login on a Mac

macOS’s SSH switch: turning it on, keys and named users, the firewall, and the errors.

Read

VS Code Remote SSH

Setup, the config entry, the errors that stop it, and the one limit no setting fixes.

Read

Termius alternatives

Eight clients ranked honestly, and the route that needs no client at all.

Read

Best AI coding agents 2026

Ten terminal agents, grouped by who they’re for, every fact checked.

Read

Tailscale stuck on Starting…

What the state means, six causes, and the fixes in cost order.

Read

Vibe coding vs traditional coding

Origin, definition, where it works, where it fails, and what the other side became.

Read

mosh vs SSH

A stream versus a state: the mechanics, the losses, and what came after.

Read

Claude plan calculator

Pro vs Max 5x vs Max 20x vs API, by hours of agent use per day

Read

tmux cheatsheet

sessions, windows, panes, and the fixes — searchable, every command verified

Read

Start

One line on any host you own.

$curl -fsSL https://spawnd.dev/install.sh | sh
Sign up free
Install the daemon

Questions

Are these guides only useful with spawnd?
No. Each one teaches the searched-for thing completely with the tools you already have, and names spawnd only where that honest route runs out. Most readers should get what they came for before the product appears.
How current are the agent facts?
Each page carries the date it was written and last checked. Agent CLIs change monthly; commands and flags are verified against the vendor’s current documentation at the date shown, and a page whose facts drift is corrected or pruned, never left to mislead.
Which agents does spawnd run?
Claude Code, Codex, OpenCode, and Aider ship as built-in shortcuts, and any CLI can be added — an agent is a named command in a real shell. If it runs in a terminal, it runs here.