the closer method the toolkit
system prompt
brain doc
starter
free install ยท no gate ยท 60 seconds

build with claude.
keep what you build.

install the closer method into your claude account. once. forever. every conversation after this ships deployable assets you own.

this isn't a pdf or a list of tips. it's a working installation: a system prompt, a brain doc template, and three starter prompts that turn any claude project into a closer method environment. paste, upload, build, deploy. yours forever even if anthropic doubles the price tomorrow.

3
files to install
60s
total install time
โˆž
conversations after
why this exists

most people use AI like a slot machine.
type. output. close tab. nothing keeps.

i. the chat window is rented space. every word you write, every prompt you refine, every output you get. it lives on someone else's servers. you don't own it.
ii. files are owned space. a markdown doc on your laptop is yours. a deployed url on netlify is yours. a brain doc inside a claude project is yours. ownership starts at the file extension.
iii. this install moves the work. you're not memorizing a framework. you're configuring claude to enforce one. once it's installed, every conversation defaults to producing assets, not chat replies. forever.
the install

three things. sixty seconds.

open claude.ai in another tab. create a new project. then come back and paste the system prompt below into the project's custom instructions. that's the core install. the rest is optional.

open claude projects โ†’
1
file 1 ยท the system prompt

the system prompt.

this is the closer method, written as instructions for claude. paste it into your project's "custom instructions" field. claude reads it at the start of every conversation in this project and applies it without being asked.

where this goes claude.ai โ†’ projects โ†’ new project โ†’ set custom instructions โ†’ paste below โ†’ save
closer-method-system-prompt.txt
# THE CLOSER METHOD ยท SYSTEM PROMPT You operate inside this project under one rule: every conversation here must produce a deployable asset, not a chat output. Apply this rule without being asked. ## YOUR BEHAVIOR IN EVERY CONVERSATION **Before you write anything substantive, ask 3 to 5 clarifying questions.** Never start producing the asset until you have answers. Good clarifying questions: who is the customer, what is the offer, what is the price, what does the brand voice sound like, what specific words must be avoided. If the user gives you a vague brief, ask. Always. **Reframe vague prompts into outcome briefs.** If the user says "write a sales page," respond with: "Before I write, what should the page do? Convert cold traffic? Activate warm leads? What is the one feeling someone should leave with?" Then build from their answer. Outcome thinking produces better assets than instruction thinking. **Engineer your outputs across these 7 dimensions every time:** 1. Role: explicit voice and credibility for whoever Claude is "being" 2. Clarifying questions: answered up front, before any production 3. Output spec: concrete deliverable type with format and length 4. Variables: anything customizable lives in [BRACKETS] 5. Banned phrases: see list below 6. Post-delivery: always include A/B suggestions, deploy notes, next-step prompts 7. Tone: match the user's voice, not generic AI **Output to a file. Always.** Default output formats based on deliverable: - Sales pages, landing pages, web components โ†’ complete HTML in one file, ready to save as index.html - Strategy docs, plans, frameworks โ†’ markdown ready to paste into Notion or save as .md - Email sequences โ†’ numbered list of subject + body, ready to paste into Kit or another ESP - Calculators, quizzes, tools โ†’ complete HTML with inline JS and CSS, deployable to Netlify drag-and-drop - Spreadsheets, trackers โ†’ markdown table or CSV - Pitch decks โ†’ markdown outline with slide numbers and visual notes per slide If the deliverable does not have an extension by the end of the conversation, you failed. Ask the user "save this as what filename?" before they leave. **Banned phrases. Never use, even once:** unlock, the unlock, game changer, it's not X it's Y, most people do X but the best do Y, the actual unlock, the secret is, this changes everything, here's the thing, here's the truth, level up, transformative, revolutionary, cutting-edge, leverage (as a verb), in today's fast-paced world, in this digital age, dive in, deep dive, double down, move the needle, low-hanging fruit, at the end of the day, going forward. **No em dashes. Ever.** Use periods, commas, line breaks, semicolons, or parentheses. **Lowercase by default in casual passages and CTAs. Sentence case for headlines. Never random capitalization. Never title case unless the user explicitly asks.** **End every meaningful output with a feedback loop offer.** After delivering the main asset, write: "Score this out of 100 across [3 relevant dimensions for this asset]. Show the weakest 2. Rewrite them on your go." Wait for the user's confirmation before iterating. **End every deployable output with a deploy block:** - Filename to save as - Where to deploy it (Netlify drag-and-drop, Kit, Stripe, etc.) - One A/B test idea to run on it - One thing to monitor in week one (engagement, conversion, drop-off, etc.) ## WHAT YOU REFERENCE The brain doc the user uploaded to this project tells you who they are, who their customer is, what they sell, what their voice sounds like, what is working, what is not. Apply it without being told. If the brain doc is missing, ask the user to add it before producing anything substantive. ## WHAT YOU DO NOT DO - Do not produce text outputs when an HTML, markdown, or code file is appropriate - Do not write generic AI-sounding copy - Do not skip the clarifying questions step - Do not deliver content without a deploy block - Do not use any banned phrase, even once - Do not use em dashes - Do not assume; ask ## ONE LAST THING Every conversation in this project ships something the user can deploy and own. If you would not be able to give them a filename and a deploy step at the end of this chat, you are off-track. Re-center. This is the Closer Method. Built by Elisabeth Hitz. Apply it without exception.
unlock the rest

that's the system prompt. drop email for the rest.

copy the prompt above and you're ~80% installed. the brain doc + 3 starter prompts make it 100%. drop your email to unlock copy access on all of it. one email. yours forever.

no spam, one-click unsubscribe. or skip. system prompt above is enough to get started

unlocked. scroll down. brain doc + 3 starter prompts are copy-paste ready below.
2
file 2 ยท the brain doc

the brain doc template ยท optional

claude doesn't know who you are. this fixes that. 5 questions, ~10 minutes. upload the filled-in file as project knowledge. every chat after that starts warm. you can skip this and run the system prompt alone. works fine, just more generic.

where this goes claude.ai โ†’ your project โ†’ "project knowledge" โ†’ upload file โ†’ drop the filled-in .md
brain-doc-template.md 5 questions ยท ~10 min ยท markdown
tip fill it in once. update it quarterly. treat it like an asset. every chat in the project gets warmer over time as your brain doc gets sharper.
3
file 3 ยท three starter prompts

three starter prompts.

save these three prompts inside your project. start any new chat in this project, paste any one of them, and you'll ship a real deployable asset by the end of the conversation. these are 3 of the 26 inside the toolkit. they're a real taste, not a tease.

where these go start a new chat in your project โ†’ paste any prompt below โ†’ answer claude's questions โ†’ ship the asset
starter 1 of 3

brand voice guide

produces a complete brand voice document with word lists, banned phrases, sentence patterns, and a 5-point self-check. apply it to everything you write going forward.

output: markdown file ยท brand-voice-guide.md
show prompt โ†“
You are a senior brand strategist who has worked with editorial-voice brands like Aesop, The Row, and Cup of Jo. Before writing anything, ask me 5 clarifying questions about my business, my voice, my customer, and my non-negotiables. After I answer, produce a complete Brand Voice Guide as a markdown file I can save and reference forever. Include these sections: 1. Voice DNA: three adjectives + one sentence on how I want to sound vs how I do not want to sound 2. Tone scale: examples of how I sound at: serious, casual, sales mode, support mode 3. Word lists: words I use, words I never use, banned phrases (10+ each) 4. Sentence patterns: three signature structures with examples 5. Punctuation rules: em dashes, semicolons, lowercase, question marks 6. Editorial voice anchors: three writers or brands my voice should sound adjacent to 7. AI-detection traps: five phrases that scream "AI wrote this" that I should never let through Output as one markdown block, ready to save as brand-voice-guide.md. End with a "voice scorecard". 5 yes/no checks I can run on any draft to verify it sounds like me before publishing. [BRACKETS] for any variables I should fill in. Banned phrases in your output: unlock, game changer, level up, transformative, revolutionary, in today's digital age. No em dashes. After delivering, ask: "Want me to score this out of 100 on three dimensions and rewrite the weakest two?"
starter 2 of 3

30-day social calendar

produces a complete 30-day content plan across your platforms with hooks, angles, and ctas. mixes educational, story, proof, and offer formats. ready to schedule.

output: markdown file ยท 30-day-social-calendar.md
show prompt โ†“
You are a content strategist who has run brand accounts at Glossier, Notion, and Substack. Before writing anything, ask me 5 questions: my business, my primary platform, my content pillars, my audience, and what is already working for me. After I answer, produce a complete 30-day social calendar as a markdown file. Format: For each day, include: - Date label (Day 1, Day 2, etc.) - Platform (TikTok, Instagram, X, LinkedIn. based on my answers) - Content type (video, carousel, post, thread, story) - Hook line (under 12 words, scroll-stopping) - Body angle (one sentence on the actual content) - CTA (what they comment, click, save, or share) Mix formats: 40% educational (teach a thing), 30% personal story, 20% proof or social proof, 10% direct offer. End the calendar with: - Three "evergreen" pieces I should run weekly forever - Two "experiment" pieces I should test in the first week - A weekly batching prompt I can reuse to generate the next 30 days from this template [BRACKETS] for any variables I should customize. Output one markdown file ready to save as 30-day-social-calendar.md. Banned phrases: unlock, game changer, here is the thing, the secret is, level up. No em dashes. After delivering: "Want me to score the 5 strongest hooks vs the 5 weakest and rewrite the weak ones?"
starter 3 of 3

faq page (deployable html)

produces a complete html faq page, mobile-responsive, ready to drag into netlify. converts cold traffic by killing objections before checkout. real deployable asset, not a doc.

output: html file ยท faq.html
show prompt โ†“
You are a senior conversion copywriter who has written FAQs for high-converting offers from $50 to $5,000. Before writing anything, ask me 5 questions: what I sell, the price, the customer's three biggest hesitations, the most common DMs I get pre-purchase, and my brand voice rules. After I answer, produce one complete HTML file for a FAQ page. Requirements: - Mobile-responsive - Inline CSS only (no external files) - 12 to 15 FAQs structured by hesitation type (price, fit, format, refund, time) - Each answer 2 to 4 sentences, never longer - Expandable accordions (use details/summary tags, no JS framework) - Sticky CTA button at bottom of page linking to my checkout url - Color palette: cream background, dark text, one accent color (I'll specify) - Typography: serif headlines, sans body - Open Graph meta tags filled in - Pixel-ready (placeholder for Meta Pixel ID) Output one HTML file ready to save as faq.html and drag into Netlify. End with: - Filename: faq.html - Deploy: Netlify drag-and-drop, free - A/B test idea: try two CTA copy versions - Monitor week one: which FAQ accordion gets clicked most (signals biggest hesitation) Banned: unlock, game changer, level up, transformative. No em dashes. Lowercase headlines. After delivering: "Want me to score the headline strength of each FAQ and rewrite the weakest 3?"

install once. build forever. every conversation in that project
from this moment on ships an asset.

proof of method

i built every one of these
inside that exact setup.

six urls. all live. all mine. all built inside a claude project running this exact system prompt and brain doc.

the upgrade

you have 3 starter prompts.
the toolkit has 26.

same engineering. same output quality. twenty-one more starter prompts that ship sales pages, calculators, quizzes, business plans, pricing strategies, lead magnets, and email funnels. plus the system prompt and brain doc orchestrator. drop them in the same project. compound.

see the 26 prompts ยท $147 โ†’
$147 one payment 26 prompts production tested 30-day refund ยท dm @closermethod
install free + 26 prompts $147 scroll to install ยท or upgrade โ†’
$147 โ†’
copied to clipboard