Frequently Asked Questions
Frequently asked questions
37 answers · More in the documentation
How can I see all my AI coding tools’ usage limits in one place on a Mac?
Install UsageNow and click its symbol in the menu bar. One popover lists every tool you track — Codex, Claude Code, Antigravity, Gemini CLI, Kiro, Warp, OpenCode, Qoder, Qwen Code, Cline, Grok Build, DeepSeek, Kimi, OpenRouter, Ollama Cloud, Windsurf, Cursor, GitHub Copilot, Perplexity, ElevenLabs, Z.ai, and JetBrains AI — with each limit window, how much is left, and when it resets, plus today’s tokens and requests.
How do I check my Codex usage limits?
With Codex installed and signed in on your Mac, UsageNow shows your Codex 5-hour and weekly limits, reset times, and plan in the menu bar. It asks the official Codex app-server, so it never reads your Codex credentials.
How do I see Claude Code usage limits?
Turn on Settings › General › Fetch Claude usage limits in UsageNow. It’s experimental and off by default; UsageNow then reads your existing Claude Code sign-in from the keychain, keeps the token in memory only, and shows your 5-hour and weekly limits. Tokens, requests, and plan appear without it.
How do I check my Antigravity quota without opening its usage panel?
Antigravity without the IDE. If the IDE cannot answer, read the official agy /usage report from compatible CLI 1.x versions (1.1.11 or later). No agent turn, no credential access by UsageNow; reports are cached for five minutes. Verified with CLI 1.2.2.
How do I see how many Kiro credits are left this month?
Turn Kiro on in UsageNow. It reads the monthly limit Kiro itself records on your Mac and shows the share left, the reset date, and your plan, plus the credits and prompts you used today. It makes no request to Kiro and reads no Kiro credential.
How do I track Warp agent usage on a Mac?
UsageNow reads Warp’s own database, read-only, and shows today’s agent requests, the credits those conversations spent, and the tokens each model used. Your prompts and the agent’s replies are never read. Warp keeps no current credit limit on the Mac, so no percentage is shown.
How do I track OpenCode usage and cost?
Turn OpenCode on in UsageNow. It reads OpenCode’s own database on your Mac, read-only, and shows today’s tokens, requests, and models, what they would cost at API prices, and a chart of the last 30 days. Only the numbers are read — never your prompts or the replies.
How do I see how many credits Qoder spent?
Turn Qoder on in UsageNow. It adds up the credits Qoder charged for each answer — from the transcripts Qoder writes on your Mac — for today and each of the last 30 days. Qoder’s sign-in is never opened. Your plan’s remaining credits aren’t shown, because Qoder keeps them only in its app.
Can I see my Ollama Cloud limits in the menu bar?
Yes, experimentally. Add a Cloud API key and UsageNow shows how much of your session and weekly allowances is left. Ollama doesn’t document this, so it may stop working without notice; if it does, UsageNow says the limits are unavailable rather than guessing. The key stays in your Keychain and goes only to ollama.com.
Can I see my AI coding usage over the last 30 days?
Under today’s numbers, each provider draws its last 30 days: one bar per day, today the brightest, with the month’s tokens or credits, what they would cost at API prices, and the model you used most. Hover a bar to see that day. Quota pace. A local chart and a clearly labelled estimate for the most critical window, based on at least three fresh quota observations spanning 15 minutes. Cached repeats, stale readings, long gaps, resets and unknown window lengths do not produce a forecast. Estimates never use token counts.
Can I see my DeepSeek, Kimi, or OpenRouter balance in the menu bar?
Yes. Turn the provider on, add your own API key, and UsageNow shows what the service reports: DeepSeek’s and Kimi’s balance, and OpenRouter’s spending today and this month with what’s left on the key. The key is kept in the Keychain on your Mac, sent only to that provider, and deleted when you turn the provider off.
How much would my AI coding tokens cost at API prices?
UsageNow shows an estimate beside each provider’s totals: what today’s tokens would cost at published API prices, with cached tokens priced as cached. It’s a sense of scale, not a bill — a subscription doesn’t charge per token.
Where can I download UsageNow?
From usagenow.com, as a signed and notarized disk image, or from GitHub Releases. You can also build it from source with Xcode.
How do I know the download is genuine?
The app is signed with a Developer ID and notarized by Apple, so macOS checks it when you open it. To check the file itself, run shasum -a 256 on UsageNow-1.0.0.dmg and compare the result with 2ecae9d0b754651e6621646689be7e7c5835f685ea793256d55644f8020bad98.
What do I need to run it?
A Mac with macOS 15 or later — UsageNow runs natively on both Apple silicon and Intel. Install at least one supported tool — Codex, Claude Code, Antigravity, Gemini CLI, Kiro, Warp, OpenCode, Qoder, Qwen Code, Cline, Grok Build, DeepSeek, Kimi, OpenRouter, Ollama Cloud, Windsurf, Cursor, GitHub Copilot, Perplexity, ElevenLabs, Z.ai, or JetBrains AI — and use it once, so there is something to show.
Is UsageNow free and open source?
Yes. It’s free, and the source code is on GitHub under the MIT License.
Do I need an account?
No. UsageNow has no accounts and no server of its own.
Is there a Windows or Linux version?
No. UsageNow is a native macOS app.
Is there a command-line tool?
Optional CLI: usagenow status and usagenow status --json read the app’s safe snapshot, with freshness and reset times. No provider runtime, network, authentication or history export. Install from Settings › General; shell configuration is left to the user.
How do I report a problem?
Choose Settings › About › Report a Problem. UsageNow shows the whole report first — versions, screen sizes, settings, and what each provider shows, never names, file paths, keys, or prompts — then opens a prefilled GitHub issue or email for you to send.
Does UsageNow update itself?
Yes, from 0.4.0. It checks for and installs its own updates, and only ones signed by UsageNow. A check asks usagenow.com for the update feed and sends nothing about you or your Mac. Coming from 0.3.0 or earlier, install 0.4.0 by hand once.
Does UsageNow upload my prompts?
No. UsageNow never stores or uploads prompts, responses, or conversation content.
Does UsageNow read my source code?
No. From session files it reads only timestamps, identifiers, model names, token counts, and rate limits. Lines with conversation content are skipped.
Does UsageNow modify my Claude Code credentials?
No. UsageNow never refreshes, modifies, or stores your Claude Code credentials.
Does UsageNow read my Codex credentials?
No. Codex’s own app-server signs in with Codex’s credentials; UsageNow only asks it for your rate limits.
Does UsageNow read my Google sign-in for Gemini CLI or Antigravity?
No. For Gemini CLI it checks only whether a sign-in exists and never opens credential files. For Antigravity it asks the running app’s own server on 127.0.0.1 for the limits it already shows, and sends nothing beyond your Mac.
Is it safe to give UsageNow an API key?
The key stays in the Keychain on your Mac and is sent over HTTPS only to that provider’s own host, only to read your balance. Redirects aren’t followed and an answer from any other host is refused, so the key can’t be passed on. It is never logged, never in the widget or analytics, and deleted when you turn the provider off. These services don’t offer read-only keys, so create a separate key for UsageNow where you can.
Does the app send analytics?
Only if you turn it on. Sharing anonymous usage analytics is opt-in and off by default; while it’s off, the app sends nothing. When it’s on, events carry only the app and macOS version, the architecture, and a random installation ID — never usage numbers, model names, prompts, code, or accounts. This website counts page views with cookieless, self-hosted analytics.
What does “58% left” mean?
58% of that usage window remains, so 42% is used. That’s the default. To see what’s used instead, choose Settings › General › Show limits as › Used — the popover, the menu bar, and the widget then read “42% used”.
Why don’t token counts match my plan’s usage?
Token and request counts are local activity recorded on this Mac. They include cached input and don’t include other devices, so they don’t necessarily match quota consumption or billing.
Why do different tools show different kinds of limits?
Each tool has its own quota system. Codex and Claude Code have 5-hour and weekly windows, Antigravity shares limits across groups of models, Kiro meters monthly credits, and Gemini CLI, Warp, OpenCode, and Qoder report none. UsageNow shows each exactly as reported, so percentages from different tools aren’t directly comparable.
Why does Claude sometimes show “Usage limits unavailable”?
Claude Code doesn’t expose quota through a supported local API, and the optional experimental source may be turned off or may not have a current sign-in. Local activity still appears.
Why doesn’t Gemini CLI show usage limits?
Gemini CLI doesn’t record usage limits on your Mac, and its quota service needs your Google sign-in, which UsageNow won’t use. You still see today’s tokens, requests, and models.
Can I use Antigravity without running the IDE?
Antigravity without the IDE. If the IDE cannot answer, read the official agy /usage report from compatible CLI 1.x versions (1.1.11 or later). No agent turn, no credential access by UsageNow; reports are cached for five minutes. Verified with CLI 1.2.2.
Which tools does UsageNow support?
Codex, Claude Code, Antigravity, Gemini CLI, Kiro, Warp, OpenCode, Qoder, Qwen Code, Cline, Grok Build, DeepSeek, Kimi, OpenRouter, Ollama Cloud, Windsurf, Cursor, GitHub Copilot, Perplexity, ElevenLabs, Z.ai, and JetBrains AI. For each one: Codex — usage limits, reset times, plan, and activity by model; Claude Code — activity by model and plan, with experimental, opt-in usage limits; Antigravity — limits and reset times from the IDE or compatible official CLI; Gemini CLI — tokens, requests, and models, today and over 30 days — Gemini CLI reports no usage limits; Kiro — the monthly credits left on your plan, when they reset, and credits used today and over 30 days; Warp — today’s agent requests, credits, and tokens by model; OpenCode — tokens, requests, models, and estimated cost, today and over 30 days; Qoder — credits spent today and over 30 days, and the answers they paid for; Qwen Code — tokens, requests, models, and estimated cost, today and over 30 days; Cline — tokens, API requests, models, and the cost Cline recorded, today and over 30 days; Grok Build — tokens, model calls, models, and the cost xAI reported, today and over 30 days; DeepSeek — the balance left on your API account, in its own currency; Kimi — the balance left on your API account; OpenRouter — what you spent today and this month, and what’s left on your key; Ollama Cloud — your session and weekly limits; Windsurf — daily and weekly limits, reset times, or billing-cycle credits from Windsurf’s local cache; Cursor — reported personal plan allowance — no shared team pool; GitHub Copilot — reported allowances and credit counters, without invented percentages for unlimited limits; Perplexity — web credit balances and reported allowances — not Sonar API usage; ElevenLabs — monthly characters, allowance, and the reset date when reported; Z.ai — global Coding Plan quota windows and reported MCP allowance; JetBrains AI — local AI Assistant quota and refill times — experimental; active-subscription validation pending.
Does UsageNow support GitHub Copilot or Cursor?
Yes, since 0.8.0. Both are experimental, explicit connections: Cursor uses a web session you connect; Copilot uses your own GitHub token with an internal Copilot endpoint, not the official billing API or VS Code’s identity. Availability depends on your plan and credential.
Will UsageNow support more providers?
New sources are added only when their data can be read reliably with clear permission. No dates — things ship when they’re ready.