THE INTRO CARD is free — what an AI skill card is and how to use one Download →Lost Lauren, contact us

THE INSTRUMENTS · Nº 978 · EDITION 1 · LAUNCHING SOON

LUND CARDS · THE INSTRUMENTSNº 978 PROPOSED
THE TOKEN COUNTER CARD — the LSC-1 face
THE CARD AS TEXT · FOR EYES AND FOR MACHINES

THE TOKEN COUNTER CARD · Counts the tokens you have used in Claude Code: your total since the first session, and where each session's went.

WAKES WHEN Someone asks how many tokens they have used · A usage limit is reached too soon · Someone asks which connector or helper costs most · Someone wants on The Token War board

THE LOOP

  1. Find the session with `find`.
  2. Run `session`, adding `--extra` for separate runs.
  3. Read the verdict and the groups.
  4. Draw one chart.
  5. Explain in a few plain bullets, with the one change that saves most.

NEVER

  • Reads someone else's sessions
  • Turns tokens into dollars
  • Calls an estimate exact

IT MAKES session — split one session's usage into groups; find — list session transcripts, newest first; all — add up every session on this computer; board — write a Grove Wars entry for THE TOKEN WAR.

THE STANDARD per-call totals are read; days the record does not hold are never filled in; cache reads; a session's own meter is read; only the person's own sessions; tokens are never turned into dollars; an entry is never uploaded for anyone; verify by execution; say the limit out loud in the work; nothing hidden; the person decides.

IN THE BOX CARD.md · CHANGELOG.md · LICENSE-NOTE.md · MANIFEST.md · MY-LEDGER.md · SELFTEST.md · SKILL.md · START-HERE.md · TIN.md · WITNESS.md · card-machine.svg · card.html · card.json · card.svg · doors/HOSTS.md · doors/local/Modelfile · doors/mcp/mcp_card.py · doors/rules/.cursorrules · doors/rules/AGENTS.md · doors/rules/CLAUDE.md · doors/rules/GEMINI.md · doors/rules/copilot-instructions.md · doors/system/CARD.system-8k.txt · doors/system/CARD.system.txt · references/THE-GROUPS.md · scripts/tokens.py · tests/EVALS.md · tests/evals.json.

PROOF SELFTEST 25/25 · witnessed 2026-10-14 on material and live on Claude · Edition 1.

THE SEAL 96 of 100 · LEGENDARY · WITNESSED · Trigger 5 · Machinery 5 · Law 5 · Portability 4 · Proof 5

IT CAN BE WRONG an AI reading a text file: it can be wrong and can invent a number, a rule, a name or a date; check what matters against your own records and the source; the risk is yours. LICENSE-NOTE.md rides in the packet and is the licence.

THE PUNCH · SHA-256 OF SKILL.MD · FA6D72610CD1C508377BBD42

THE FACE, FULL SIZE → · THE MACHINE FACE → · card.json →

THE TOKEN COUNTER CARD

Counts the tokens you have used in Claude Code: your total since the first session, and where each session's went.

SESSION · FIND · ALL · BOARD

96LEGENDARYWITNESSEDIN CLAUDEIN LUNDRIN CHATGPT

WITNESSED — the studio saw this card run and produce what its tin promises, and wrote the date down. What the marks mean →

THE SPEC SHEET · READ BEFORE YOU BUY · THE STANDARD EVERY CARD MEETS →

SEAL
96 of 100 · LEGENDARY · WITNESSED
EDITION
E1 · witnessed 2026-10-14
MACHINE
scripts/tokens.py · 4 commands · selftest 25/25
RUNS
On your own computer. Standard library only, offline; no account, no key, no telemetry.
HOSTS
Claude · Claude Code · ChatGPT · Gemini · any AI that reads a file (HOSTS.md names each door)
LICENCE
One person, any number of their own AIs and computers · editions replace files · nothing renews · freeze, not loss
PUNCH
FA6D72610CD1C508377BBD42 · sha-256 of SKILL.md · check any file against it →

CITE THIS CARDTHE TOKEN COUNTER CARD, Edition 1. Lund Studio LLC, Westminster, Colorado. https://lundr.tech/?view=token-counter-card. Punch FA6D72610CD1C508377BBD42. Retrieved 2026-10-15.

What it is.

Counts how many tokens you have used in Claude Code: your total since the first session or since any day (since January 1), by day, month and project, or for one session what used them (instructions, subagents, each connector, web research, file work), read from the session transcripts on this computer and weighted the way tokens really count.

Use when someone asks how many tokens they have used, their total usage, what ate their usage limit, which connector or helper costs most, or to enter The Token War on the Grove Wars board. Do not use to weigh a skill before install (THE LOAD CARD), to count tokens in a document (THE SOUNDING CARD), or to read anyone else's sessions.

It carries a machine that shows its working: session — split one session's usage into groups; find — list session transcripts, newest first; all — add up every session on this computer; board — write a Grove Wars entry for THE TOKEN WAR — 25 checks in its selftest. It will never reads someone else's sessions; never turns tokens into dollars; never calls an estimate exact.

Eleven laws, written into the card.

They are in the card itself, so the AI follows them.

PER-CALL TOTALS ARE READPer-call totals are read, never guessed; the split between groups is labelled an estimate.
DAYS THE RECORD DOES NOT HOLD ARE NEVER FILLED INDays the record does not hold are never filled in; the report says where the record starts.
CACHE READSCache reads, cache writes and output are weighted as they really count.
A SESSION'S OWN METER IS READA session's own meter is read: what it holds beyond the transcript is added and marked, never counted twice.
ONLY THE PERSON'S OWN SESSIONSOnly the person's own sessions.
TOKENS ARE NEVER TURNED INTO DOLLARSTokens are never turned into dollars.
AN ENTRY IS NEVER UPLOADED FOR ANYONEAn entry is never uploaded for anyone: the person sends it, after agreeing to the board's terms.
VERIFY BY EXECUTIONVerify by execution, not by reasoning: run the machine and read what it prints before saying a thing is done.
SAY THE LIMIT OUT LOUD IN THE WORKSay the limit out loud in the work: every result names what it did not check.
NOTHING HIDDENNothing hidden: two readers, the person and the AI, see the same words.
THE PERSON DECIDESThe person decides.

What rides in the card.

In the card’s own voice, with a machine that shows its working.

THE GUIDEreferences/THE-GROUPS.md
THE MACHINEtokens.py · four commandssession — split one session's usage into groups; find — list session transcripts, newest first; all — add up every session on this computer; board — write a Grove Wars entry for THE TOKEN WAR. Every command prints its working and refuses a bad input; 25 checks in its selftest.

What it does once your AI has it.

  • session Split one session's usage into groups
  • find List session transcripts, newest first
  • all Add up every session on this computer
  • board Write a Grove Wars entry for THE TOKEN WAR
  • On the record Live on Claude Sonnet, 2026-10-14: 13 of 13 passed; the card fired on 8 of 8 trigger cases and on none of its hand-offs. The session meter equals the usage records Claude Code keeps, field for field, in 325 of 328 runs checked.

What it will never do.

It will never reads someone else's sessions; never turns tokens into dollars; never calls an estimate exact.

It sees only the transcript files on this computer. It cannot see claude.ai chats, other devices or the account's total, and its split between groups is an estimate. It is an AI reading a text file: it can be wrong, so check what matters against your own records and the source.

What’s in the buy.

  • CARD.md — the whole card: its boot, laws, loop, when it says no, its seals, ledger, term and score · SKILL.md — the door your AI reads first · START-HERE.md — the first five minutes, for you
  • THE-GROUPS.md — the guide in references/, in the card's own voice
  • scripts/tokens.py — the machine: session · find · all · board, standard library only, no network, selftest 25/25
  • WITNESS.md — the runs on real material and live on Claude, with what was not run · SELFTEST.md — the machine's own test, as run at the cut · MANIFEST.md, every file hashed · CHANGELOG.md
  • tests/evals.json + EVALS.md — the cases it must take, hand off and refuse
  • MY-LEDGER.md — the card's memory, on your own machine · TIN.md, the nine lines on the tin · LICENSE-NOTE.md
  • card.svg + card-machine.svg + card.json · card.html, the card as one offline file · doors/ — the rules files for Cursor, Codex, Gemini, Copilot and Claude Code, the system text for any other AI, a local-model file and an MCP door

PROOF · SELFTEST 25/25 · WITNESSED ON MATERIAL · LIVE ON CLAUDE SONNET 2026-10-14 · 13 OF 13 PASSED · EDITION 1 · PUNCH FA6D72610CD1C508377BBD42

ON EVERY DOOR · A CARD FOR YOUR AI, A FILE FOR YOUR COMPUTER

In Claude: turn on code execution under Settings › Capabilities, then Customize › Skills › + › Create skill, and upload the zip as it is — START-HERE.md in the box walks you through it, and tells you how to keep the ledger. In ChatGPT, Gemini and the rest: a Project, a GPT or a Gem with the card’s files attached and the short door in the box pasted as its instructions. The words run anywhere an AI reads Markdown; the machines run wherever the host runs Python. Witnessed in Claude only. The host marks →

Take it.

Nº 978 PROPOSED — LAUNCHING SOON

Sold as is under the term printed. A card instructs an AI you operate; results depend on your model, your files and your judgement. Not medical, legal, tax or financial advice, never the person’s boss, and not a licence to practise. The terms →

ONE PAYMENT · NOTHING RENEWS ITSELF

OFFLINE · ONE LICENCE · PUNCH-CHECKABLE · THE STANDARD

One payment. The file is yours to keep and keeps working at its edition; new editions issued in the year after you buy reach you by email as replacement files. Tax is added at checkout where it applies. Nothing auto-renews.

One email when it is on sale, from the studio; nothing else, ever.

A TERM, NOT A SUBSCRIPTION: NOTHING RENEWS ITSELF AND NOBODY IS CHARGED AGAIN. WHEN A TERM ENDS THE FILES STAY ON YOUR COMPUTER AND STAY READABLE — FREEZE, NOT LOSS.

THE MACHINE · IN THE OPEN · FROM THE CARD’S OWN WITNESS.MD

tokens.py, with its record in the open.

A small Python program, standard library only. Your AI runs it; you can too. The blocks below are the record from the card’s own WITNESS.md, as printed: the session meter checked against the usage Claude Code records, and the live runs on Claude.

sessionsplit one session's usage into groups
findlist session transcripts, newest first
alladd up every session on this computer
boardwrite a Grove Wars entry for THE TOKEN WAR
selftest25 checks that prove the machine; it must print 25/25
THE METER, CHECKED AGAINST COMPLETE RECORDS
- **The meter reads true.** On this computer, 328 separate `claude -p` runs each kept a complete transcript and a meter. In 325 of them the meter equals the transcript's own usage records field for field (fresh input, cache reads, cache writes, output) for the model that did the work. Their meters also carry the small side calls Claude Code makes with a faster model, which never appear in a transcript (313,220 raw tokens in all). The other 3 are files that two runs wrote to, where the meter holds only the last run; there the transcript is larger and nothing is added.
- **What the meter found.** The session that built THE MARKETPLACE CUT and Grove Wars ran from 2026-10-12 to 2026-10-15 and was compacted more than once. Its meter read 276,896,456 raw tokens; its own transcript and its helper agents' transcripts still showed 107.8 million. Edition 3 counted the 107.8 million; Edition 4 adds the 169.1 million the transcript lost.
- **The entry against the full count.** `board --since 2026-01-01` and `all --since 2026-01-01` from the cut packet, over the same 362 transcripts: 305,156,441 raw tokens over 1,425 calls, each. The Grove Wars app's own entry checker accepted the entry (effective matches its raw parts on every day; tokens per call within its limits).
THE LIVE RUNS ON CLAUDE SONNET, 2026-10-14
Claude Sonnet, 2026-10-14, through `claude -p`, with this edition (punch see CUT-REPORT.md) installed alone beside the host's own skills.

| case | kind | card fired | answer passed |
|---|---|---|---|
| trig-1 | trigger | yes | yes |
| trig-2 | trigger | yes | yes |
| trig-3 | trigger | yes | yes |
| trig-4 | trigger | yes | yes |
| trig-5 | trigger | yes | yes |
| trig-6 | trigger | yes | yes |
| trig-7 | trigger | yes | yes |
| trig-8 | trigger | yes | yes |
| hand-1 | handoff | no | yes |
| hand-2 | handoff | no | yes |
| hand-3 | handoff | no | yes |
| ref-1 | refuse | no | yes |
| ref-2 | refuse | no | yes |

13 of 13 passed; the card fired on 8 of 8 trigger cases and on none of its hand-offs. On trig-8 (a long, compacted session's total) it answered 63.7M effective tokens (306,282,992 raw) for the computer, said the count includes the part the compactions dropped, read from the session meters, and said plainly that it had not isolated the one session. A first run of the same 13 cases, on the cut before the report's meter sentence was reworded, also passed 13 of 13 (kept as live-e4-first.jsonl).

Edition 3 (punch EED781800AEBD31E465031F3) passed 12 of 12; Edition 2 passed 11 of 11; Edition 1 passed 9 of 9 with an end-to-end run. Their records are kept in the kit.

THE RUNS ARE PASTED FROM THE CARD’S OWN WITNESS.MD AS PRINTED. FIXTURES MARKED FICTIONAL ARE THE STUDIO’S OWN EXAMPLES, INVENTED FOR THE RECORD; NO PRIVATE PERSON IS NAMED.

QUESTIONS AT THE TABLE

Asked before the wire.

What does it need from me?

Your own Claude Code session transcripts on your computer. Find the session with `find`, and it goes from there.

Does it send anything anywhere?

No. The machine is a standard-library Python program that runs on your own computer and sends nothing anywhere; the words run wherever an AI reads Markdown. No server, no account, no key.

What is on the record?

The self-test (25/25), the session meter checked digit for digit against the usage records Claude Code keeps (325 of 328 runs equal field for field), and the live runs on Claude Sonnet (13 of 13 passed; the card fired on 8 of 8 trigger cases and on none of its hand-offs) pasted into WITNESS.md as printed; the punch on the face is the SHA-256 of SKILL.md, so a downloaded copy can be checked against this room.

Can it be wrong?

Yes — it is an AI reading a text file. It sees only the transcript files on this computer. It cannot see claude.ai chats, other devices or the account's total, and its split between groups is an estimate.

What will it never do?

It will never reads someone else's sessions; never turns tokens into dollars; never calls an estimate exact.

← THE INSTRUMENTS · THE FLOOR →

AT YOUR ELBOW, NEVER THE BOSS · NOT LEGAL, MEDICAL, TAX OR FINANCIAL ADVICE · NOT A PAYMENT CARD · © 2026 LUND STUDIO LLC

PAY BYAmerican Express Apple Pay Diners ClubDiscoverGoogle Pay Mastercard PayPalShop Pay Visa
CART · 0