Completely free — every feature, no paid tier, no card

Know exactly how much AI coding capacity you have left.

AI Monitor is your AI developer usage center — usage, limits, reset timers, cost, and forecasts for Claude Code and (soon) every AI provider you use, in one place. On your desktop, phone, watch and home screen.

Sign in with GitHub or Google — no password to create. Prefer your phone? Download the Android app (APK, not on Play Store yet).

See it before you sign in

Real screenshots from the running apps, using a demo account — not mockups, and not anyone's actual usage.

The full picture — usage, trend, history and rankings in your browser.

AI Monitor's web dashboard showing the 5-hour usage window, today's cost and burn rate
Dashboard — where you stand right now
AI Monitor's analytics page showing tokens and cost per day
Analytics — how usage is moving
AI Monitor's stats page showing an activity heatmap, streaks and session counts
Stats — streaks, sessions and an activity heatmap
AI Monitor's leaderboard showing ranked handles, levels and badges
Leaderboard — opt-in, ranked under a handle

Android and Wear OS screenshots are coming — the apps are real and downloadable today, but these shots have to be taken on a physical device.

Claude Code dashboard

Example data — not a live account
5-hour limit
Resets in 1 hr 47 min82%
Last updated: 2 minutes agoLive82 / 100
Weekly · all models
Resets in 3 days 14 hr61%
Last updated: 2 minutes agoLive61 / 100

Today

Requests
184
Tokens
1.82M
Estimated cost
$4.82

Cost is always labeled estimated — never presented as an invoice.

One dashboard, every provider

Claude Code 82%
OpenAI Codex (coming soon)47%
Gemini (coming soon)31%

Accurate, never fabricated

If a provider doesn't expose a metric through a supported mechanism, we show "Unavailable from provider" — never a guess.

Local-first Claude integration

Claude Code usage is derived from session data already on your machine. We never ask for your Anthropic password.

Burn-rate forecasting

See current, average, and peak burn rate, plus an estimated time to your limit — only when there's enough history to trust it.

Frequently asked questions

Does this read my Claude conversations or source code?

No. We only read timestamp and token-count metadata that Claude Code already records locally — never message content, file contents, or tool inputs/outputs.

How is estimated cost calculated?

Token counts multiplied by published Anthropic pricing. It's always labeled "estimated" — never presented as an actual bill.

What if a metric isn't available?

We show "Unavailable from provider" rather than inventing a number. See our Security and Docs pages for details.

Know where your usage stands, in two minutes

Every feature, free. No paid tier, no trial, no credit card — nothing to cancel later.

Sign in with GitHub or Google