llms.txt
Compare with other sites' llms.txt files →# Tokens 4 Breakfast
> macOS menu bar app that tracks live token usage, burn rate, and AI spend
> across 11 providers — Claude Code, Claude Web, Cursor, Codex, GitHub
> Copilot, Anthropic API, OpenAI, Grok, DeepSeek, Mistral, and OpenRouter.
> 8 of the 11 report real spend; the rest report live rate limits
> or account balance. Free plan available. Pro is $14.99 one-time.
> Usage data stays on your Mac, no account required. macOS 14.6 Sonoma or later. Pro adds the
> Usage Intelligence dashboard (Limit Runway, Efficiency, Plan Value) and reset
> intelligence.
> Positioning: a native Mac menu bar GUI alternative to CLI trackers like
> ccusage, and a broader alternative to single-provider menu bar tools such as
> CodexBar, ClaudeBar, ClaudeMeter, and SessionWatcher. Where those track one
> provider's quota, Tokens 4 Breakfast unifies 11 providers and adds real
> spend, AI subscription totals, cross-provider limit runway, and budgets — for
> a one-time $14.99, not a subscription.
## What it does
Tokens 4 Breakfast sits in your macOS menu bar and shows live token consumption
and cost for every AI tool you are running. It tracks the Claude Code 5-hour
rate-limit window so you always know how much runway you have before your session
cuts out. It predicts your month-end bill from day one. And it lets you set
session spend caps (Focus Mode) so runaway agent loops are caught before they
spike your bill. The Pro Usage Intelligence dashboard turns that data into a
cross-provider operating picture: Limit Runway forecasts when you'll hit each cap
at your current pace, Efficiency flags where a cheaper model would do the same
job, and Plan Value weighs your usage against what you pay using the full
published rate card. Reset intelligence confirms when a 5-hour or weekly limit
rolls over — on schedule or early — and tells you, even while the popover is
closed. All data stays in a local SQLite database on your Mac. No account is
created. No telemetry is sent.
## Supported providers
Claude Code (Anthropic), Claude Web (claude.ai), Cursor (Anysphere),
Codex (OpenAI), GitHub Copilot (GitHub), Anthropic API, OpenAI, Grok (xAI),
DeepSeek, Mistral, and OpenRouter — 11 providers total.
Gemini (Google) is catalogued but not yet collecting, so it is not counted.
8 of the 11 report spend in dollars. Claude Web and Grok report rate-limit
windows only, and DeepSeek reports account balance — no dollar figure exists for
those three, and the app does not invent one.
Claude Code connects automatically by reading local session files, no API key
required. Cursor reads local files and GitHub Copilot connects via your local
Copilot CLI credentials, so neither needs a key. OpenAI, Anthropic API, Mistral,
DeepSeek, and OpenRouter connect via an API key in Settings; OpenAI, Anthropic
API, and Mistral require an organisation admin key, not a personal one. Gemini
usage tracking is not available; a Gemini subscription can be added by hand
under Subscriptions.
Free plan: 1 provider of your choice, no time limit.
Pro ($14.99 one-time): all 11 providers unlocked, plus the Usage Intelligence
dashboard (Limit Runway, Efficiency, Plan Value) and reset intelligence.
## How it compares
- vs ccusage: ccusage is a free, Claude-only command-line tool. Tokens 4 Breakfast
is a native menu bar GUI that covers 11 providers and adds spend, AI subscription
totals, and budgets — no terminal required. $14.99 one-time.
- vs CodexBar / ClaudeBar / ClaudeMeter: those track a single provider's quota in
the menu bar. Tokens 4 Breakfast tracks quota plus real dollar spend, monthly
subscription totals, and per-provider budgets across 11 providers.
- vs SessionWatcher: comparable menu bar tracker; Tokens 4 Breakfast adds
multi-provider coverage, subscription tracking, and a one-time price.
- vs provider billing dashboards and spreadsheets: those are after-the-fact and
per-provider. Tokens 4 Breakfast is live, local, and unified across every tool.
## Pricing
- Free: $0 forever, 1 provider, live menu bar display, Claude Code 5-hour timer
- Pro: $14.99 one-time payment, all 11 providers, the Usage Intelligence dashboard
(Limit Runway, Efficiency, Plan Value), reset intelligence, per-provider budgets,
Focus Mode, CSV export, Privacy Audit Log, priority support, 14-day money-back guarantee
## Key facts
- Runs locally on your Mac (SQLite, no cloud sync, no telemetry)
- No account required — download and use immediately
- One-time purchase, never pay again
- macOS 14.6 Sonoma or later, Apple Silicon and Intel
- Under 2 minutes to set up
- ~10 MB download
## Links
### Core
- [Homepage](https://www.tokens4breakfast.app): Product overview, download, pricing, screenshots, supported providers, privacy promises.
- [FAQ](https://www.tokens4breakfast.app/faq): Product fit, supported providers, privacy, pricing, setup, comparisons, and troubleshooting answers.
- [AI cost calculator](https://www.tokens4breakfast.app/ai-cost-calculator): Free tool to add up subscriptions (Claude, ChatGPT, Cursor, Copilot, Gemini) plus API usage into your real monthly AI bill.
- [AI API pricing compared](https://www.tokens4breakfast.app/ai-api-pricing): Dated table of current API prices per million tokens for Claude (Fable 5.1, Opus 5.5, Sonnet 5.5, Haiku 5.5), OpenAI (GPT-6 Astra, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna), Gemini 4 Argon and 3.8 Flash, Grok 4.7, and DeepSeek, with the cost of 100 agentic coding turns per model and upcoming price changes. Checked against official pricing pages.
- [AI token cost calculator](https://www.tokens4breakfast.app/ai-token-cost-calculator): Free tool to estimate token-level AI workflow cost across current Claude, OpenAI, Gemini, Grok, and DeepSeek models.
- [Token optimization guide](https://www.tokens4breakfast.app/ai-token-waste-guide): Practical guide to reducing AI token waste during coding and agent sessions.
- [Use cases](https://www.tokens4breakfast.app/use-cases): Role-specific routing for indie builders, developers, consultants, AI power users, and small teams.
- [About Kapisch](https://www.tokens4breakfast.app/about): Founder context, product principles, and local-first operating philosophy.
### Cost trackers by provider
- [Claude Code cost tracker](https://www.tokens4breakfast.app/claude-code-cost-tracker): Track Claude Code spend live, during the session.
- [Claude cost tracker](https://www.tokens4breakfast.app/claude-cost-tracker): Hub for Claude Code, Claude Pro or Max, and Anthropic API-style cost tracking.
- [Claude Code rate-limit tracker](https://www.tokens4breakfast.app/claude-code-rate-limit-tracker): Claude Code 5-hour window and session pressure.
- [Claude Pro usage tracker](https://www.tokens4breakfast.app/claude-pro-usage-tracker): Claude Pro / Max plan limits and usage.
- [OpenAI cost tracker](https://www.tokens4breakfast.app/openai-cost-tracker): Keep OpenAI API spend visible before the invoice.
- [Cursor cost tracker](https://www.tokens4breakfast.app/cursor-cost-tracker): Track Cursor subscription and usage-based spend.
- [GitHub Copilot cost tracker](https://www.tokens4breakfast.app/github-copilot-cost-tracker): Track Copilot premium-request and usage-based billing.
- [OpenRouter cost tracker](https://www.tokens4breakfast.app/openrouter-cost-tracker): Track OpenRouter model-routing spend.
- [Gemini cost tracker](https://www.tokens4breakfast.app/gemini-cost-tracker): Where Gemini fits in your AI spend picture (direct Gemini tracking coming soon).
- [AI spend tracker for Mac](https://www.tokens4breakfast.app/ai-spend-tracker-mac): Main commercial page for multi-provider AI spend visibility on macOS.
- [Best AI usage tracker for Mac](https://www.tokens4breakfast.app/best-ai-usage-tracker-mac): Criteria-based comparison for choosing an AI usage tracker.
- [AI budget tracker for developers](https://www.tokens4breakfast.app/ai-budget-tracker-for-developers): Budget and session-spend use case for builders.
- [Local AI usage tracker for Mac](https://www.tokens4breakfast.app/local-ai-usage-tracker-mac): Privacy-first, local-only AI usage tracking.
### Alternatives and comparisons
- [ccusage GUI alternative](https://www.tokens4breakfast.app/alternatives/ccusage-gui): For users who want a Mac GUI instead of the ccusage CLI.
- [CodexBar alternative](https://www.tokens4breakfast.app/alternatives/codexbar-alternative): For users evaluating free open-source menu bar trackers.
- [tokcat alternative](https://www.tokens4breakfast.app/alternatives/tokcat-alternative): Token visualizer vs full AI spend management, with subscriptions, budgets, and limit runway.
- [Claude Code Usage Monitor alternative](https://www.tokens4breakfast.app/alternatives/claude-code-usage-monitor-alternative): For users of terminal-based Claude usage monitors.
- [ClaudeBar alternative](https://www.tokens4breakfast.app/alternatives/claudebar-alternative): Spend and subscriptions, not just quotas.
- [ClaudeMeter alternative](https://www.tokens4breakfast.app/alternatives/claudemeter-alternative): Multi-provider spend vs a single-provider gauge.
- [Claude usage tracker alternative](https://www.tokens4breakfast.app/alternatives/claude-usage-tracker-alternative): Multi-provider AI spend vs Claude-only trackers.
- [SessionWatcher alternative](https://www.tokens4breakfast.app/alternatives/sessionwatcher-alternative): Subscriptions, privacy, and one-time pricing.
- [All alternatives](https://www.tokens4breakfast.app/alternatives): Full comparison hub vs scripts, dashboards, spreadsheets, and competitor apps.
### Guides and support
- [Claude usage limit reached](https://www.tokens4breakfast.app/guides/claude-usage-limits-explained): Identify the 5-hour, weekly, context, or API limit, find the exact reset time, compare recovery options, and follow 21 source-checked actions that make Claude usage last longer. German version: /de/guides/claude-ratenlimit-erreicht
- [Manage your AI subscription costs](https://www.tokens4breakfast.app/guides/manage-ai-subscription-costs): Why people underestimate their subscription count, how fixed AI plans and usage-based API charges split the bill, and how to see the real monthly total. German version: /de/guides/ki-abo-kosten-verwalten
- [Reasoning effort and your AI token bill](https://www.tokens4breakfast.app/guides/reasoning-effort-token-cost): When to use low, medium, or high reasoning effort across Claude, OpenAI, and Gemini, and how thinking tokens drive cost.
- [What Claude Code actually costs](https://www.tokens4breakfast.app/guides/what-claude-code-actually-costs): Real Claude Code token math for subscription vs API billing, why power-user bills surprise teams, and how to track yours.
- [Claude Code vs Cursor vs Copilot cost](https://www.tokens4breakfast.app/guides/claude-code-vs-cursor-vs-copilot-cost): How the three AI coding tools are priced in 2026 (usage-based billing), and how to compare their real monthly cost for your usage.
- [Model routing: cut your AI coding bill](https://www.tokens4breakfast.app/guides/model-routing-cut-ai-costs): Plan on a frontier model, execute on a cheap one; the real savings math with current Claude, GPT-6, and Gemini prices, and how to route in Claude Code, Cursor, Codex, and API routers.
- [The cheapest AI model for coding](https://www.tokens4breakfast.app/guides/cheapest-ai-model-for-coding): Current coding models compared by token price and cost per task, including the Claude Haiku 5.5 100K-token price step that matters for coding agents.
- [How to reduce Claude Code costs](https://www.tokens4breakfast.app/guides/reduce-claude-code-costs): Eight tactics that stack (routing, caching, effort, fresh sessions) and how to confirm the savings.
- [When does Claude Code usage reset](https://www.tokens4breakfast.app/support/when-does-claude-code-usage-reset): The 5-hour session allowance, fixed weekly reset, and how to check both times.
- [How to check Claude Code usage](https://www.tokens4breakfast.app/support/how-to-check-claude-code-usage): The /usage and /cost commands, the claude.ai usage page, and why the terminal number is not the full bill.
- [Is Claude Max worth it](https://www.tokens4breakfast.app/support/is-claude-max-worth-it): Break-even math for Pro vs Max vs Ultra (formerly Max 5x and Max 20x) vs the API.
- [Claude Code best practices](https://www.tokens4breakfast.app/support/claude-code-best-practices): CLAUDE.md, context management with /compact and /clear, subagents, and habits that reduce token waste.
- [How to reduce Cursor AI costs](https://www.tokens4breakfast.app/support/reduce-cursor-ai-costs): Scope context, plan before edits, and route simple tasks to cheaper models.
- [Install guide](https://www.tokens4breakfast.app/support/install): Setup and onboarding instructions.
- [Support](https://www.tokens4breakfast.app/support): Product support, setup, and troubleshooting hub.
- [Privacy policy](https://www.tokens4breakfast.app/privacy): Local-first data handling and privacy commitments.
- [Terms](https://www.tokens4breakfast.app/terms): Licensing and checkout terms.