HOW THE MACHINE WORKS. WHY YOU CAN TRUST THE OUTPUT.
GENERATION LOGIC · MODEL TRANSPARENCY · SATISFACTION COMMITMENT · FIELD REPORTS
HOW EACH PRODUCT IS BUILT
No PDFs sit on a server waiting to be sold. Each product is generated fresh on purchase. Here is the exact path from checkout to delivery.
Stripe verifies the payment and fires a webhook to the machine. The product slug, your email, and the responsible unit (NEO, ROCKY, RIPLEY, or ORPHEUS) are passed in.
The machine pulls the product metadata: title, table of contents, target operator, voice constraints, and the unit's system prompt. This becomes the brief.
The brief is sent to the Anthropic Claude API. Claude writes the full playbook in the unit's voice, following the documented structure. No templates. No reused paragraphs.
The output is rendered into a typeset PDF and emailed to you. Total time from purchase to inbox: typically under 60 seconds.
The system prompt and generation logic are documented on the Autopilot page. You can see exactly how the machine operates before you buy anything.
WHAT CLAUDE SONNET IS — AND WHY IT'S RELIABLE
Claude Sonnet is a frontier large language model built by Anthropic, an AI safety company founded by former OpenAI researchers. It is one of the most capable text generation models currently available, used by enterprise teams at Notion, Quora, DuckDuckGo, and the U.S. government.
Three reasons the output is reliable for the kind of work this catalogue does:
- Long context. Claude can hold the entire system prompt, product brief, and unit voice constraints in a single pass without losing the thread. It doesn't drift mid-document.
- Instruction adherence. Claude is trained to follow specific structural and tonal directives. When the brief says "operator-to-operator, no hype, lead with mechanism" — that is what comes out.
- Honest reasoning. Anthropic's training process penalizes hedging filler and generic advice. Claude is meaningfully better than alternatives at producing concrete, specific recommendations instead of vague platitudes.
This is why every product reads like it was written by a specialist, not assembled from templates. The model is doing the work, and the model is good.
NOT USEFUL IN 7 DAYS? WE MAKE IT RIGHT.
Every purchase is backed by a 7-day satisfaction commitment. The terms are short on purpose.
- If a product does not deliver clear, actionable value within 7 days of purchase, email agent@aimachina.shop.
- State the order ID and one sentence on what didn't land.
- The machine will either re-generate the playbook against a tighter brief, replace it with a more relevant unit's product, or refund the purchase in full. Operator's choice.
- No process. No argument. No retention questions. The machine stands behind its output.
Applies to every product in the catalogue. Commission and operator subscriptions follow the same principle, scoped to their respective deliverables.
OPERATOR RESULTS — GROWING IN REAL TIME
Real results submitted by active operators. This log is appended as reports come in — it is not a curated marketing wall. Format: OPERATOR · DATE · PRODUCT PURCHASED · RESULT REPORTED.
// LOG STATUS: OPEN · APPENDS IN REAL TIME AS REPORTS ARE RECEIVED
Operators: send results to agent@aimachina.shop with the product purchased and a one-line outcome. Verified entries are added to this log.