AgentPort vs MeterTab

Agent Port is a local-first monitor and remote control surface for the AI coding agents you already run — Claude Code, Codex, and other CLI agents living in tmux on your Mac. Check on them, steer them, and reply to prompts from your iPhone, anywhere on your own network. WHY AGENT PORT Long-running agents don't finish on your schedule. Agent Port keeps you in the loop without keeping you at your desk: see at a glance which sessions are running, which are waiting for your input, and which are done — then jump in and respond. WHAT YOU CAN DO • Live session list — every agent pane with a clear status: running, waiting for you, done, or idle • Real-time logs — stream a session's output as it happens, with copy and auto-follow • Full terminal — a real PTY bridge when you need to type anything, not just presets • Quick replies & keys — send text, common answers, and control keys with one tap • Token usage — total Claude Code and Codex token consumption, right at the top • Multiple machines — save server profiles and switch between your Macs instantly • Picture-in-Picture — keep a session's log floating while you do other things (iPhone) LOCAL-FIRST & PRIVATE There is no Agent Port cloud. The app talks directly to a small, free, open-source service you run on your own Mac, over your LAN, Tailscale, or any network you control. Your code, sessions, and tokens never pass through our servers — because there are none. The app collects no analytics and stores its settings only on your device. TRY IT INSTANTLY On first launch the app opens in Demo mode with fully populated sample data — no server, no account, nothing to configure. When you're ready for real use, run the open-source service on your Mac and switch profiles with one tap. Source code, setup guide, and documentation: github.com/xiaoxiunique/agent-port Requires a Mac running the free Agent Port service to monitor your own agent sessions.

MeterTab puts your AI coding budget on your iPhone, Apple Watch, and Home Screen. See at a glance how much of your Claude, Codex, and Cursor budget you have left - before you hit a wall mid-task. WHY METERTAB AI coding assistants burn through 5-hour and weekly limits with no warning. MeterTab turns the usage the tools already log into a live budget you can plan around - synced from the MeterTab Mac app to your iPhone and Watch over your own private iCloud. WHAT YOU GET • Live budget - 5-hour and weekly usage with time-to-reset countdowns. • Answer your agents from anywhere - approve, deny, or pick an option when Claude Code asks, from your Mac, iPhone or Watch. • See what every session is doing - which worktrees are busy, the command in flight, and the agent's own task list. • iPhone & Apple Watch apps, Home Screen & Lock Screen widgets, and a Watch complication - your budget on every wrist and screen. • Multi-provider - Claude Code, Codex, and Cursor in one place. • Claude service status - on the main screen, in Settings, and on your widgets and complication, so an outage never looks like your budget running out. • Per-model cost breakdown - see exactly which models are spending your budget. • Activity heatmap + burn rate - spot your heaviest hours and how fast you are trending. • Every number carries its source - Claude's own figure, measured, or estimated - so you always know what you are reading. • Pay-as-you-go spend too - the API costs your Mac tracks from your own provider keys land here as well. • Speaks your language: English, German, Spanish, French, Italian, Japanese, Korean, Dutch, Portuguese (Brazil) and Chinese (Simplified) - with figures formatted for your region. FOR AGENTS: A BUDGET THEY CAN TALK TO MeterTab's Mac app runs a local MCP server, so your own AI agents can ask "how much budget is left?" and throttle themselves - downshift to a cheaper model, or pause - instead of failing halfway through a job. Your iPhone and Watch keep that same budget in your pocket. PRIVATE BY DESIGN No backend. No account required. No data collected. Any keys you give the Mac app stay in its Keychain, on that Mac. Your usage never leaves your devices - iCloud sync uses your own private iCloud database and carries only counts. FREE & PRO Free: the live gauge, threshold notifications, and up to 4 providers. Pro (one-time purchase, covers Mac + iPhone + Watch via Universal Purchase, with a 14-day free trial): the MCP server your agents can query, billing-key exact-dollar spend tracking, iCloud sync, the heatmap, burn rate, unlimited providers, and the iPhone & Watch apps. A NOTE ON ACCURACY Anthropic does not publish exact token limits, so the 5-hour and weekly percentages are estimates: your Mac works them out from its Claude Code logs, and they can differ a few percent from Claude's own usage screen. Calibrating on the Mac anchors them and carries across to here. If you connect your Claude account there, your iPhone and Watch show the exact figures too - every number is tagged so you can tell at a glance which kind you are looking at. MeterTab is an independent tool and is not affiliated with, endorsed by, or sponsored by Anthropic, OpenAI, or Cursor. Product names are used only to describe compatibility.