THE DEPLOY CARD · a bot on the web, its key never in a browser
WAKES WHEN someone wants their own bot on the web, or a bot folder is about to be uploaded
THE LOOP
page --name · the chat page, AI said
worker --provider --model · the server
put the key in as a secret, never a file
check <folder> · SAFE TO DEPLOY?
deploy; set the origin to the page
NEVER
Put a key in a page, a file or a message
Let a page call the provider directly
Leave the origin open to everyone
IT MAKES page — index.html, the chat page; worker — worker.js and wrangler.toml; check — keys, direct calls, open origins: SAFE TO DEPLOY or not.
THE STANDARD the key is a secret; the page calls the owner's endpoint only; one origin; size is capped; the page says it is an AI; the machine sends nothing.
PROOF SELFTEST 7/7 · witnessed in Claude 2026-10-10 · Edition 1.
THE SEAL 98 of 100 · LEGENDARY · WITNESSED · Trigger 5 · Machinery 5 · Law 5 · Portability 5 · 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. THE-CLAUSES.md rides in the packet and is part of the licence.
THE PUNCH · SHA-256 OF SKILL.MD · BC6D012B122223243422F2FE
CITE THIS CARDTHE DEPLOY CARD, Edition 1. Lund Studio LLC, Westminster, Colorado. https://lundr.tech/?view=deploy-card. Punch BC6D012B122223243422F2FE. Retrieved 2026-10-10.
What it is.
Puts a bot on the web without its key ever reaching a browser: page writes the chat page, which talks only to the owner's own endpoint and says on its face that it is an AI; worker writes the small server — a Cloudflare Worker and its wrangler.toml — that holds the key as a secret, answers only the owner's page origin, caps the size of what it takes and forwards it with the owner's system prompt to Anthropic or OpenAI; check reads the whole folder before it goes up and names anything that looks like a key, a page that calls the provider directly, a key written into the settings, or an origin left open — SAFE TO DEPLOY, or not, without ever printing the key it found.
It carries a machine that shows its working: page — index.html, the chat page; worker — worker.js and wrangler.toml; check — keys, direct calls, open origins: SAFE TO DEPLOY or not — 7 checks in its selftest, every command on the record. It will never put a key in a page, a file or a message; never let a page call the provider directly; never leave the origin open to everyone.
Six laws, written into the card.
They are in the card itself, so the AI follows them.
THE KEY IS A SECRETIt goes in with wrangler secret put API_KEY; it is never written into a page, a settings file or a message, and check never prints one it finds.
THE PAGE CALLS THE OWNER'S ENDPOINT ONLYA page that calls the provider would need the key in the browser; check names it.
ONE ORIGINThe worker answers only the page it was made for; an empty or open origin is named.
SIZE IS CAPPEDTwenty turns and about twenty-four thousand characters a request; the rest is refused before it costs anything.
THE PAGE SAYS IT IS AN AIThe line is on the page from the first write: it can be wrong, and nothing private should be shared.
THE MACHINE SENDS NOTHINGIt writes files and reads them; the worker it writes reaches the provider only once the owner deploys it.
What rides in the card.
In the card’s own voice, with a machine that shows its working.
THE GUIDEreferences/THE-HORIZON.md
THE MACHINEdeploy.py · three commandspage — index.html, the chat page; worker — worker.js and wrangler.toml; check — keys, direct calls, open origins: SAFE TO DEPLOY or not. Every command prints its working and refuses a bad input; 7 checks in its selftest, every command on the record.
What it does once your AI has it.
page Index.html, the chat page
worker Worker.js and wrangler.toml
check Keys, direct calls, open origins: SAFE TO DEPLOY or not
On the record The chat page written, talking only to its own endpoint and saying it is an AI; the worker and its settings written for one origin, with the key as a secret; the folder read SAFE TO DEPLOY; then a key pasted into the settings was named and the folder read NOT SAFE.
What it will never do.
It will never put a key in a page, a file or a message; never let a page call the provider directly; never leave the origin open to everyone.
The worker is a starting server: it caps size and origin, but rate limits per visitor, abuse filters and logging are the owner's to add for a public bot. The key check reads the common shapes of a key; a key in another shape can pass, so keep keys out of folders as a habit. The worker it writes does reach the model provider once deployed — that is its job — and costs what the provider charges. 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: who it is, its laws, when it says no, what it carries · SKILL.md — the door your AI reads first · START-HERE.md — the first five minutes, for you
THE-HORIZON.md — the guide in references/, in the card's own voice
scripts/deploy.py — the machine: page · worker · check, standard library only, selftest 7/7
WITNESS.md + SELFTEST.md — every command on the record, as it printed · MANIFEST.md, every file hashed · AUDIT.md, the twelve checks
MY-LEDGER.md — the card's memory, on your own machine · TIN.md · LICENSE-NOTE.md · THE-CLAUSES.md · HOSTS.md
card.svg + card-machine.svg + card.json · card.html, the card as one offline file · the doors for ChatGPT, Gemini, Cursor, Claude Code and MCP hosts (rules/, system/, mcp/)
PROOF · SELFTEST 7/7 · 3 COMMANDS ON THE RECORD · WITNESSED IN CLAUDE 2026-10-10 · EDITION 1 · PUNCH BC6D012B122223243422F2FE
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 1108 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. 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.
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 WITNESSED RUN OF 2026-10-10
deploy.py, every command showing its working.
A small Python program, standard library only. Your AI runs it; you can too. Every block below is pasted from the card’s own WITNESS.md, as printed.
page
index.html, the chat page
worker
worker.js and wrangler.toml
check
keys, direct calls, open origins: SAFE TO DEPLOY or not
selftest
7 checks that prove the machine; it must print 7/7
page · INDEX.HTML, THE CHAT PAGE
$ python3 scripts/deploy.py page --name The Bakery Bot --out site
written: site/index.html · calls /api/chat · says it is an AI
worker · WORKER.JS AND WRANGLER.TOML
$ python3 scripts/deploy.py worker --provider anthropic --model model-id-here --system system.md --origin https://bakery.example.com --out site
written: worker.js and wrangler.toml · anthropic · the key goes in as a secret: wrangler secret put API_KEY
check · KEYS, DIRECT CALLS, OPEN ORIGINS: SAFE TO DEPLOY OR NOT
$ python3 scripts/deploy.py check site
3 files read · SAFE TO DEPLOY
(someone pastes the key into wrangler.toml as a plain setting)
$ (someone pastes the key into wrangler.toml as a plain setting)
check · KEYS, DIRECT CALLS, OPEN ORIGINS: SAFE TO DEPLOY OR NOT
$ python3 scripts/deploy.py check site
KEY wrangler.toml sets API_KEY as a plain variable — make it a secret
3 files read · NOT SAFE TO DEPLOY · 1 to mend
THE NUMBERS, NAMES AND DATES IN THESE RUNS ARE THE STUDIO’S OWN EXAMPLES, INVENTED FOR THE RECORD; NO REAL COMPANY, PERSON OR PROPERTY IS NAMED.
QUESTIONS AT THE TABLE
Asked before the wire.
What does it need from me?
Page --name · the chat page, AI said. It asks first for the bot's name, the provider and model, the system prompt, and the page's origin.
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 (7/7) and every command run on real example inputs, pasted into WITNESS.md as they 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. The worker is a starting server: it caps size and origin, but rate limits per visitor, abuse filters and logging are the owner's to add for a public bot. The key check reads the common shapes of a key; a key in another shape can pass, so keep keys out of folders as a habit. The worker it writes does reach the model provider once deployed — that is its job — and costs what the provider charges.
What will it never do?
It will never put a key in a page, a file or a message; never let a page call the provider directly; never leave the origin open to everyone.
