Get everything out of Moshi
Deep dives that thread Moshi's features into a real workflow — running coding agents, keeping a durable workspace, and steering it all from your phone. Each guide goes past the basics into the patterns that make remote agent work feel effortless.
Moshi with Claude Code
The end-to-end playbook: prep the host, connect, approve from the lock screen and watch, prompt by voice, review diffs, watch your rate limits, and keep agents running while you're away.
Moshi with Codex
Codex-specific from end to end: pick an approval mode that matches your phone, teach the repo with AGENTS.md, tune reasoning effort, answer sandbox escalations from the lock screen, and resume with one command.
Running a fleet of agents
Patterns for driving several agents at once without losing track of which one needs you.
Moshi with tmux
Turn tmux into a durable, phone-shaped workspace your agents and shells live in.
Moshi with Herdr
An agent-aware multiplexer on the host, an agent-aware terminal in your pocket — workspaces in the picker, taps instead of chords, and events that land on the exact tab.
tmux or Herdr?
The honest decision: tmux's ubiquity, launcher, and copy-mode depth versus Herdr's agent state, workspace picker, and tab-level deep links.
Moshi with an always-on Mac
Make a spare or desk Mac into an endless agent machine you can drive from your phone.
Moshi with Tailscale
The clean way to reach home machines: a tailnet instead of port forwarding, Easy Pair that detects it automatically, and the Tailscale SSH fork explained properly.
Hands-free with voice
Prompt and steer agents by voice — the engines, the two modes, and a real hands-free loop.