This takes you from "I got an invite email" to "I'm building pages for clients." No technical background needed, and you'll never have to touch a scary terminal. Read it top to bottom — each step builds on the last.
Here's the whole mental model. You've been using Claude in a browser chat. What we use is the Claude Code app — same Claude, but it runs on your computer and can actually do things: set itself up, build pages, organize files.
Everything happens inside that one app. You type into its message box the same way you'd text a coworker — in plain English. When you see a command in this guide (like /startwork), you just type it into that same box and hit enter. You never open a separate "Terminal" program. If a command looks technical, don't worry about understanding it — paste it in and Claude runs it for you.
If you're ever unsure, just ask Claude in plain English. "What does this mean?" "What should I do next?" "Can you set that up for me?" You cannot break anything by asking, and it's almost always the fastest way through.
Do this on day one, in order. Brad is around all morning, so if a step doesn't work, say so immediately instead of pushing through it. A broken step is our bug, not yours.
Sign into your christian@landersrx.com inbox. You'll have an email from HQ inviting you to LandersRX. Open it and keep it handy — it has a one-time invite link you'll use in a couple steps. Don't share it; it's tied to you.
This is a different app from the Claude chat you've used. Go to claude.ai/code, download the desktop app for your Mac or PC, install it, and sign in with your Claude account. Open it up — that's your workspace from here on.
Already pay for Claude? Same login works.
In the Claude Code message box, type the line below and hit enter. Claude does the heavy lifting — it'll ask permission as it goes, and you just say yes.
Set up HQ for me by running: npx create-hqIf anything pops up that you don't recognize, type "what should I do here?" and let Claude guide you. When it finishes, type /setup to run the quick welcome wizard.
This is the one step where people hit friction. If it stalls, errors, or asks for something you don't have, stop and message Brad rather than trying to force it. Getting stuck here is expected and it is a five-minute fix on our end. Do not lose an afternoon to it.
This connects you to our shared workspace. Type the command below and sign in using christian@landersrx.com — it has to match the address the invite went to.
/hq-loginCopy the invite link from that Step 1 email, then type the two commands below (paste your link where shown). This downloads the full LandersRX workspace — clients, brand kits, playbooks, and the skills that build pages.
/accept <paste-your-invite-link-here> /hq-syncWhen this finishes, you're officially on the team and everything's on your machine.
This is your "I'm sitting down to work" command. Run it and Claude shows you what's on your plate.
/startworkCheck off all five: /startwork runs · /hq-sync finishes clean · you can open Motion, Figma, Replo and Slack · you've said hi in Slack · you know what you're doing tomorrow. Anything not checked, tell Brad today, not next week.
Four things to know and you'll feel at home fast. None of this is technical — it's just how to drive the app.
Watch this instead — 11 min →Brad walking through all of it on screen: staying in the HQ folder, running /hq-sync each morning, one session per task so you don't burn your usage, and the two commands that matter — /startwork and /handoff. Watch it before you read the rest of this section. Everything below is the same material written down, so you have it to refer back to.
Press Shift + Tab to switch how much Claude checks with you. You'll see the current mode at the bottom of the screen. Three to know:
Claude figures out a plan and shows it to you before doing anything. Nothing changes until you approve. Use this whenever you're starting something new — it's the safe default.
Claude does the work without stopping to ask each step. Faster. Switch to this once you've approved a plan and just want it to run.
Claude asks permission before each action. Good when you want to watch closely. This is what you start in.
Simple rule: start new work in Plan mode, flip to Auto-accept once you like the plan.
A session is a single conversation. When you start something unrelated, begin a fresh session (or type /clear) so old context doesn't confuse things. Don't live in one endless chat — fresh and focused works far better.
Type /startwork whenever you begin — each time you sit down, and again when you switch to a different project. It catches you up on where things stand. Type /handoff when you finish a chunk of work or stop — it saves everything so your next /startwork picks up right where you left off.
Think per task, not per day: new project? /handoff to save → /clear → /startwork fresh. And /handoff again whenever you're done for the day.
Now and then Claude asks "Do you want to allow this?" That's a good thing, not an error. Read it, and if it's what you asked for, say yes.
Everything above, boiled down to a single card you can keep open or print. Pin it next to Claude Code for your first week.
Open the quick-reference cheat sheet →Everything else is detail. Start with the video — it is the shortest route to why we do this. The two documents under it are the how.
Why we build ad-dedicated pages — 8 min →The core premise of the company, from Jordan and Brad. We are not a CRO shop: we build net new pages that continue the story of a brand's top-performing ads, matched to a customer persona and its value props. Treat the page as the second half of the ad. Includes what that has actually produced — roughly 60% more scale on one client by sending a persona to its own page, and about 2x better new-customer efficiency on another. If you only watch one thing, watch this.
The Playbook — how LandersRX actually works →What we sell, who does what, and the full delivery workflow from access granted to pages live. Read sections 1 through 6 in order. Skim the rest as reference.
The Quality Bar — what good looks like →Every way we've gotten a page wrong for a client, turned into rules. Never fabricate social proof. Never crop a model's head. Put our product first in a comparison chart. Roughly 90 of them, grouped so you can actually learn them.
You are not expected to memorize it. You are expected to run the pre-flight checklist at the bottom of it before anything you make goes to a client. Every time.
You'll violate some of these in your first month. Everyone does. The list exists so you violate each one once, instead of learning it the expensive way on a live client.
Every video in one place, in the order they make sense. Two of them are linked higher up the page next to the thing they explain — they are here too so nothing is missing.
What to expect working here — 6 min →The honest version: your first weeks will feel like a fire hose, and asking for help early is the expected behaviour, not a bad signal. Also covers how we do feedback — direct, both directions — and the daily rhythm: respond, post a quick update on what you did, and run anything client-facing past Brad before it goes out.
Why we build ad-dedicated pages — 8 min →The premise of the whole company. Also linked under Read these first, above the Playbook.
The walkthroughs below follow one client from data to live page to results, and on to the next round. They are longer and denser than the two above. Watch 1 and 2 in your first week; 3 onward when you reach that step for real. One is still being recorded — the arc is shown in full so you can see where the job goes, and that card says what to do meanwhile.
How to actually drive the tooling. Also linked under Claude Code basics, which is the written version of the same material.
2 · Building a client’s page roadmap — 28 min →Where every engagement starts. Pull the account’s context out of Motion — landing page analysis and messaging theme tags over the last 30 days, sometimes 90 — then split the roadmap across products weighted by where spend actually is. Watch the top creatives by spend and purchase value and take notes on hooks, problem/solution and audience, then turn those into page ideas and an organized roadmap. The ideas start with you, not the model. AI fills the middle and a human reviews the end; skip that and you ship slop.
3 · From roadmap to page outline, part 1 — 29 min →Building the first listicle page inside HQ off an approved roadmap. Generating context from the round-one roadmap, then reviewing and reshaping the proposed outline — headline, subheadline, comparison chart, and the reasons mapped to a belief ladder. The important half is the back-and-forth: correcting claims against real customer review language, cutting the comparison chart down, catching the model’s factual mix-ups about the product, and only building once the story lines up.
3b · From feedback to a revised page, part 2 — 21 min →The same page, picked up after the design feedback lands. A brand kit that did not exist during part 1 gets generated and applied mid-stream, then the layout gets worked over edit by edit: the announcement bar and menu fold into one sticky CTA, seven reasons come down to five, headlines get bolded, the reviews and social-proof and FAQ blocks move or come out, the offer section gets restyled after the Pretty Boy pages. Mobile order is headline, image, then supporting text — that one is a rule, not a preference. Watch how much gets removed; the revision round is mostly subtraction.
The page those two videos build — live →Open this after you have watched 3 and 3b. It is the finished article for that pair — the same page, live, after the revision round. Read it for structure and copy rather than craft; the images have not been optimized yet. Worth doing the comparison deliberately: which calls made on the recording actually survived to the live page, and which quietly did not.
4 · From approved design to a built page — 14 min →What happens once a design is done. Push it into the client’s Figma folder, one page per round, and let the client comment there. When they say the comments are in, start a fresh session and have Claude pull them, apply them and check its own work. You still compare before and after side by side, and you read any reply Claude drafts before it lands on a client’s comment. Once the client approves, the build goes to Replo’s AI builder, which copies from the page’s live link. Replo credits are paid by the client, so finish the page in HQ before it goes to Replo. Then QA the build against Figma yourself: buttons, mobile and desktop sizing, buy box, fonts, colors, images.
The parts video 4 does not get to: resizing and optimizing images before the build, and the repo side of shipping a page. Ask before you attempt your first deploy.
5 · Launching pages and reading performance — 31 min →How pages go live in Meta and how you tell whether they are working. The goal you set with the client is more profitable volume, not beating the page they already run. Our preferred rollout: in a CBO, duplicate the top performing ads, leave the originals live, point the copies at the new page and let Meta decide where the money goes — spend is signal. Check the launch the day it goes live: the top ads (not a couple of one-off statics) are pointed at the pages, and the URL and display link actually changed. Expect a read in three days and a clear one by a week. Then the video walks through the Motion landing page board, reading the ad against the page for a message mismatch, Triple Whale for new visitors and bounce rate, and Clarity heatmaps. A Clarity recording is one person, so don’t build a change on it.
6 · What to test once a page works — 15 min →Where the next round comes from. We sell rounds to clients as page tests, not new pages, because one change to a working page can move it 10 to 40 percent. Start simple — hero, five reasons, offer — and mostly with listicles and sales pages. Once a page works, open the page depth and iteration menu in HQ: things to add (more reasons, reviews, an AI review summary, UGC, a guarantee, a results timeline, comparisons) and ways to change how the page sells (positioning, headline, reason order, proof, where the button goes, buying on the page). The menu isn’t ranked yet, so bring a reason for each change. The video builds a real three-cell test — control, an AI review summary, then the review summary plus a guarantee — and shows how to read the result. It ends on the landing page evaluation playbook: spend and efficiency decide whether a page works, and the other metrics tell you what to try next. Both documents are first versions, so tell Brad what’s missing.
7 · Running that test in Intelligems — 16 min →How the test you picked in 6 actually gets run, and how to read it. Intelligems is for changes to a page, not for versions that talk to a different buyer — it tells you which version does best on the same traffic, not how much spend each one could take. The setup is a split URL test: a clearly named control, a group for each variant, the live URL as the original and every visitor redirected. Judge it on profit per visitor, which means the client needs their product costs in Intelligems. Give it two weeks. The first few days swing wildly, and you’ll rarely get a clean statistical answer, so read the direction. Before calling it, break the result down: new against returning visitors, desktop against mobile, and which products people actually bought. The video shows a test that looked flat overall but was up 21% for one product and down 4% for another.
Not part of the sequence. Useful background, and one you will want open the first time you do the thing it covers.
How a client goes from signup to pages live — 10 min →The commercial shape of the business: what happens before someone signs, how the design review loop with Figma works, where deals typically stall, and how renewal dates get set.
Building pages from our template board — 8 min →Watch this before your first page build. It walks the Lander — Website Inspiration Templates board: real listicles, sales pages and PDPs from brands we work with, the wireframe flows built on top of them, and the story arc almost every winning page follows. Those structures are wired into the page-building skill, so the board is where you make the human calls — which page type, which flow, how much depth. Build from the template. Starting from scratch every time is how pages get overcomplicated.
How our ClickUp board is built — 9 min →A tour of the client work board and what each status actually means, so you can tell at a glance whether something is with us or with the client. Deliberately simple — there are no user assignments and no notifications yet.
Turning a Figma file into HTML — 2 min →Install the html.to.design plugin and go. Save this one for the first time you need it rather than watching it cold.
The goal of your first stretch is context, not output. You're not being tested. You're getting to the point where you can look at a client's ad account and have an opinion about it. Everything below builds toward that, in order.
Work through the Setup section above. Get into HQ, get into Slack, and confirm you can actually open every tool: Motion, Figma, Replo, ClickUp.
Don't move on until this is genuinely done. Half-finished setup is the single most common way a first week goes sideways, and it never announces itself — things just quietly don't work later.
Read the two documents linked above. Then come back with a written list of questions and we'll jam on them together.
No target number. Bring what you actually have. If you think something in there is wrong, say so — that's more useful than a clarifying question.
Start scanning our client accounts. You're not producing anything here. You're just building general awareness of who our clients are, what they sell, and what kind of work we do for them.
Poke around. Read old audits. Look at pages we've shipped. Ask Claude anything.
Brad will name the two accounts you'll be on first — you'll have them by day one. We're building new pages for both of them soon, so get ahead of it.
Work out what's actually going on in their ad accounts right now. Do your own read on it — your own informal audit, however you want to run it. Then look at the pages that are already live for them and how those are performing.
The output isn't a document. The output is that you and Brad can sit down and be on the same page about these two accounts.
Brad will pull up Claude and work through real client projects while you watch. This is the fastest way to learn how the work actually gets done — faster than any document, including this one.
Ask questions the whole way through. Interrupt. That's the point of it. Sessions get recorded so you can go back over them.
The goal is a full book of 8 to 10 clients within 60 days. Your rate steps up when you're carrying it, on the terms in your agreement. That's tied to the book, not the calendar — hit it in 30 days and it moves in 30 days.
Three things, and they never change:
You will never have to wonder whether you're doing well. If something isn't good enough, I'll tell you directly and specifically, while it's still cheap to fix. That's not criticism, it's the job. In exchange: when you're stuck or behind, tell me early. A problem I hear about on day three is a small problem. The same one on day fourteen is a client call.
Day one of anything new has rough edges. Nobody expects you to nail every step, and getting stuck is not a signal about you. Two things to lean on:
1. Ask Claude directly. Stuck on a step or staring at something confusing? Type it into the message box and ask. Nine times out of ten that's the fix.
2. Message Brad. If something is genuinely blocking you — a login won't go through, the invite link errors out, a step just won't run — reach out and we'll get you unblocked fast. In week one, no question is too small. Losing half a day to something I could have fixed in five minutes is the only real mistake here.