Agent Runway Download

For Claude and OpenAI subscribers

Know your runway.

Agent Runway lives in your Mac's menu bar and shows how much of your weekly AI allowance is left — paced against how long until it renews. One glance tells you whether to push hard or ease off.

Download for Mac How the colors work →

macOS 15 or later · Apple Silicon & Intel · Notarized by Apple · Free

FinderFileEditViewGo
Menu bar item: Claude with four blue cells and an hourglass, OpenAI with three green cells Tue 9:41 AM
The Runway panel: Claude 82% left, renews in 20 hours, blue; OpenAI 59% left, renews in 4 days, green, with a reset credit expiring in 3 days

Pacing, not prediction

Five cells. Four colors.
One glance.

Runway compares the allowance you have left with the time left in your weekly window. Burn faster than the clock and the cells warm up; bank a surplus and they cool to blue.

  • Blue — plenty available. At least twice the allowance the clock says you need, and 20% or more left.
  • Green — roughly on pace for the week.
  • Yellow — getting tight. Ease off to make it to renewal.
  • Red — very tight. You'll likely run short.

⏳ appears when you're blue and renewal is under 48 hours away: a nudge to use the surplus before it resets.

Try it

86% left
24h

What's inside

Everything the week needs, nothing it doesn't

Glide path

Your allowance across the week against the even-pace diagonal. Above the line, you're ahead.

Reset warnings

Surplus about to reset gets an hourglass. Reset credits are called out a week before they lapse — Runway never spends one.

Token activity

Hourly or six-hourly bars for this Mac, split into cached history, new input and output.

Morning report

A 6 AM notification with one line per provider, phone-friendly: 🟦🟦🟦🟦⬛ Claude · 86% · 24h ⏳

Copy a snapshot

Copy the panel as a picture — title, cells and all — and paste it straight into a chat.

Fault tolerant

Offline, signed out, CLI missing, enterprise plan — it says exactly that, keeps the last good reading, and retries.

Activity

Where the tokens actually go

Most of what an agent "uses" is its own history, re-read from the prompt cache on every turn. The Activity tab draws that as frosted glass, with the new input and output rising inside it — then flips to New only to show what you really added.

Read from the session logs Claude Code and Codex already keep on your Mac. Each byte is read once; after the first scan, refreshes take a fraction of a second.

Activity tab, all tokens over 24 hours: tall frosted bars with a thin solid base of new tokens; 96% cached
All tokens · 24 hours
Activity tab in light mode, new tokens only over 7 days: input bars with a darker output base
New only · 7 days
Claude

The same usage data behind Claude Code's /usage screen. When its sign-in needs refreshing, Runway runs claude -p /usage — a local command that costs zero tokens.

OpenAI

The Codex CLI's app-server, asked only for account and rate-limit reads.

Activity

The CLIs' own session logs, read in place on your Mac.

Read-only by design

It watches. It never spends.

  • Never sends a prompt or starts a session
  • Never uses a reset credit
  • Never refreshes or rewrites your sign-in — the CLIs do that themselves
  • Runs Claude Code with hooks off, so checks don't show up as sessions
  • No accounts, no analytics: it talks only to Anthropic and OpenAI, through their CLIs

Get started

Up and running in a minute

  1. 1
    Download

    Unzip AgentRunway.zip and drag Agent Runway to Applications.

  2. 2
    Have a CLI signed in

    Claude Code on a Claude plan, and/or Codex (npm i -g @openai/codex, then codex login).

  3. 3
    Glance up

    Runway checks hourly, when you open it, and updates itself quietly with Sparkle.

Download for Mac

Using Bartender or another menu bar manager? New items may start hidden — allow Agent Runway there.

Questions

Does checking use up my allowance?

No. Every check is a read. Claude's fallback, claude -p /usage, is a local command with no model call: zero tokens, zero cost. Codex is asked only for account and rate-limit data.

Why isn't it on the Mac App Store?

App Store apps must run in Apple's sandbox, which forbids exactly what Runway needs: running the claude and codex command-line tools and reading their session logs. Instead it's signed with a Developer ID and notarized by Apple, so it opens without warnings and never expires.

What leaves my Mac?

Only the usage requests the CLIs themselves make to Anthropic and OpenAI, plus a daily update check. Runway has no server, no account and no analytics. Its readings stay in ~/Library/Application Support/AgentRunway.

I'm on an enterprise or API-key plan.

There's no personal weekly allowance to pace, so Runway says so instead of guessing — and the Activity tab still shows your token burn.

Is the percentage exact?

The percentage comes straight from the provider and is authoritative. The cells round to the nearest fifth, with any allowance at all keeping one cell lit and all five reserved for exactly 100%.

Is this made by Anthropic or OpenAI?

No. Agent Runway is an independent utility and isn't affiliated with or endorsed by either company.