Outpost vs Tapproval

Outpost is an SSH client built for one thing above all: rendering your terminal exactly right. If you've connected to a server from your phone and watched your setup fall apart - broken Powerline segments, missing Nerd Font icons, boxes where emoji should be, mangled box-drawing - Outpost is the fix. It's engineered around glyph correctness, so your shell looks the way it does on your desktop. Glyph-perfect rendering. Nerd Font icons, Powerline prompts, box-drawing, wide and ambiguous-width characters, combining marks, and emoji - all correct. Secure by design. Passwords and imported keys are stored in the iOS Keychain behind Face ID. Or generate a key that never leaves the device's Secure Enclave. Host keys are pinned on first connect, with a hard warning if one ever changes. Built for real work. Keep several servers open at once and switch with a tap - scrollback survives. A keyboard bar with the keys a terminal actually needs. Copy and paste, drag to scroll, and tap-to-click in mouse-aware TUIs. Run your own setup commands automatically on connect. Make it yours. Font size, colour themes, and cursor style. No account, no tracking, nothing to sign up for. Your credentials stay on your device.

Tapproval is for Claude Code: a remote control for it on your Apple Watch. When the agent working on your computer asks permission — run this command? change this file? — the question lands on your wrist, in the session that asked, and a tap answers it. No phone out of the pocket, no walk back to the desk. Made for vibe coding: start Claude Code on a task, walk away, and keep it going from your wrist — on the sofa, on a walk, or in a meeting. WHAT IT DOES • Approve or deny in the session that asked, right under its conversation, or, for lower-risk prompts, straight from the alert • Answer Claude's questions and plans with the same choices as on your computer, picking more than one where Claude allows it • Your sessions are the first screen, live and grouped by what they need: which one needs a tap, which one is working, and the project each belongs to • Read the conversation the way your phone shows it, with bold, lists and plans laid out, and reply from the watch: "Proceed", "What's left?", "/code-review", or anything you dictate • Send a message and it reaches the session open on your Mac: its terminal and the Claude app on your phone show it, with the answer. What you type there shows on the watch too • With the app closed, the watch still buzzes for a card from a Mac, through your own iCloud • See your day: what was handled without you, and what you decided NOTHING TO CONFIGURE The watch finds your computer by itself — on your home Wi-Fi, or away from home through a private travel address. No account, no server of ours: your prompts and conversations stay on your own machines and in your own iCloud. WHAT YOU NEED A computer running Claude Code (any Mac or Linux machine) with the free, open-source helper installed — one command, then forget it. Curious first? The built-in demo shows the whole flow without any setup. Your first 100 answers are free. After that, one purchase of $5 unlocks unlimited answers for good. There is no subscription and no account. Reading sessions and the day's activity stays free, and a prompt you don't answer from the wrist still appears on your computer as usual, so nothing is ever blocked. An optional $1 coffee for the developer is there too.