AI Token Usage Tracker vs Model Usage

Know exactly what your Claude AI bots cost you. Monitor your Claude token burn rate and see exactly how much your AI workflows spend. Whether you're running Claude Code, OpenClaw, ClawBot, or your own custom AI agent — this tool helps you understand exactly how much your API-based AI workflows cost. Track your Claude API keys, keep an eye on what your bots are spending, and set alerts before costs get out of hand. Redesigned with Liquid Glass for iOS 26. Perfect for developers building with AI agents, coding assistants, or automated workflows. Screenshots are taken from the Light and Dark mode with different backgrounds and Glass Styles. FREE FEATURES • Nearly real-time (~1 hour delay) cost tracking in your menu bar • Monthly cost breakdown with token details (input, output, cache read, cache creation) • Workspace/project cost breakdown • CSV export of your usage data • Multi-currency support (USD, EUR) with live exchange rates • Configurable spend limit with native macOS notifications • Auto-refresh every 30 minutes • Dashboard Widgets PRO FEATURES (ONE-TIME PURCHASE) • Intraday Analysis — see today's spending hour by hour • Model Overview — exact costs per AI model • API Key Breakdown — track costs per individual API key • Unlimited API Keys — monitor multiple accounts at once • 7-day free trial included Your purchase applies also to Mac and Apple Vision Pro — buy once, use on all your devices. PRIVACY FIRST Your API keys are stored securely in the iOS Keychain. No data collection. No analytics. No tracking. All processing happens locally on your iOS device. Built for developers who use Claude Code or the Claude API and want to keep their costs under control. Also available on Apple Vision Pro and macOS. Requires an Anthropic Admin API key. Subscription-based usage (claude.ai, Claude Desktop) is not tracked.

Model Usage is a real time dashboard for your AI coding tool usage. See how much of your five hour window and weekly quota you have used, watch your session tick up live, and get notified before you hit a limit. HOW IT WORKS A small client runs on your computer and reads the usage files that your AI coding tools already write locally. It forwards that usage to a server address you control. The apps pair with that server over a QR code or short pair code. No account, no email, no login. LIVE WHERE YOU ALREADY LOOK - Live Activity: a glanceable card with your current session and weekly usage, visible on the lock screen and Dynamic Island. - Home Screen widgets: compact and expanded sizes that show your session, weekly totals, and rate multipliers at a glance. - Lock Screen widgets: a tiny usage indicator right next to the clock. PAIRS WITH THE REST OF YOUR DEVICES Once you are paired on one device, the same setup flows to every companion app automatically. Send a snapshot to anyone in a conversation with the Model Usage messaging app. PRIVACY BY DESIGN No analytics. No tracking. No accounts. No advertising SDKs. Your usage data goes from your computer to a server you choose, and then to your own devices. Model Usage never sees it. FEATURES - AI coding tool usage tracking, side by side - Five hour session window and weekly window - Live Activity with Dynamic Island integration - Companion apps across every size of screen - Messaging app for sharing usage snapshots - Lock Screen and Home Screen widgets in all sizes - Live rate multiplier awareness. Rate changes are reflected in real time. - Open pair code or QR flow. No account required. - Dark mode throughout SETUP IN UNDER A MINUTE 1. Install the desktop client. 2. Copy the pair code. 3. Enter it in the app. That is it. The setup guide lives at modelusage.app/setup. Model Usage is built by a solo developer and does not collect, store, or share any personal data. See modelusage.app/privacy for details.