Tokens

---

You already burn tokens all day. tokens.ci turns that into a public standing: how much you ran through today, how it splits across clients and models, and where you rank against everyone else doing the same thing.

bunx tokens-cli@latest login

iOS app

Your rank and usage on your phone, without opening a browser — share cards rendered on device, plus Home screen and Lock screen widgets for the day's tokens, your running total and your rank.

There is no sign-in. You enter a GitHub username and the app reads that public profile, so enter your own or the widgets will show someone else's usage.

Languages

The site is available in English, Simplified Chinese and Japanese. English is the default and carries no prefix, so every existing link keeps working; the other languages live under /zh-CN/… and /ja/…, and the switcher in the header moves you to the same page in the other language rather than to the home page.

The prefix is not cosmetic. HTML is cached at Cloudflare's edge by URL, so a scheme where one URL renders different bytes based on a cookie would let whichever request arrived first decide the language for everyone behind it. With a prefix in the URL, /zh-CN/leaderboard and /leaderboard are simply two documents.

To add a language:

1. Add the locale to LOCALES and LOCALE_LABELS in web/src/lib/i18n/config.ts.
2. Copy web/src/lib/i18n/messages/*/en.ts to the new locale directory under web/src/lib/i18n/messages/<namespace>/ and translate the values. Keys you leave out fall back to English at runtime, so a partial catalogue is usable.
3. Add the two prefixed entries (/<locale> and /<locale>/:path*) to the matcher in web/src/middleware.ts.

Messages are ICU: {name} for values, {count, plural, one {…} other {…}} for counts, and richText() from web/src/lib/i18n/rich.tsx when a sentence has to contain a link. Numbers and relative times go through Intl — see web/src/lib/format.ts.

The CLI is English unless you ask otherwise: --lang, then TOKENS_LANG, then language in settings.json. The system locale (LANG, LC_ALL) is never consulted, and --json output is always English. tokens --lang ja status renders one run in Japanese without changing anything stored.

Share cards follow the same rule: /api/og?locale=ja draws the Japanese card and the URL is the cache key, so each language is its own immutable entry. An absent or unknown locale renders English, which is what every card minted before this existed already asks for.

Why this repository is public

So you can check what we send.

The CLI reads the session files your AI clients already write to disk, totals them locally, and uploads only the totals — token counts, model names, client names, timestamps. Prompts, completions, file contents and paths never leave your machine.

You do not have to take that on faith. The whole pipeline is here:

• cli/tokens-core/src/sessions/ — one parser per client, showing exactly which fields are read out of each session file
• cli/tokens-core/src/aggregator.rs — how those reads become daily totals
• cli/tokens-cli/src/main.rs — the submit path; build_submit_payload is the one function that assembles the payload, for both a real submit and a dry run

tokens submit --dry-run summarizes what would be uploaded without uploading it; tokens submit --dry-run --json prints the exact JSON payload.

MCP server names are opt-in. By default no MCP information is sent. Run tokens mcp-names enable (or pass --include-mcp to one submit) to include the names — never commands, URLs, credentials or env values — of the MCP servers configured for the clients that submission covers (Claude Code ~/.claude/.mcp.json, Cursor ~/.cursor/mcp.json, Kiro ~/.kiro/settings/mcp.json, OpenCode skills). They are shown publicly on your profile. tokens mcp-names disable turns it off, and the next submit clears the stored list; tokens mcp-names status shows the setting and the files read.

If your gateway adds routing prefixes to model IDs, list those prefixes in ~/.config/tokens/settings.json (or $TOKENS_CONFIG_DIR/settings.json):

{
  "modelIdPrefixesToStrip": ["gateway-a/", "gateway-b/"]
}

This list is empty by default. Matching is case-insensitive; the longest matching prefix is removed once before the existing model normalization. Empty prefixes are ignored, and removal cannot produce an empty ID. Unlisted namespaces remain distinct. The rule affects local grouping and uploaded model IDs, unlike the presentation-only modelAliases setting. Pricing, token counts, costs, and raw session files and caches are unchanged. Subsequent submissions apply the rule to historical usage too. Use the same rules on devices that submit the same usage; changing the rules changes uploaded model identity.

We are not asking anyone to run their own copy. This repo exists to be read.

Supported clients

All 42 are detected automatically — if it is installed and has written sessions, it is counted.

This is the beginning of the README. Read the rest at the original. Read the full README ↗