Using Manus Till I Earn $1,000,000 LIVE

Published March 3, 2026 | By marcus

Training Notes · Build In Public Series

How I Use Manus AI To Make $100,000+ Online — Full Session Notes

Live walkthrough by Marcus — the exact workflow, tools, and business frameworks used to generate real income with Manus AI. Full session covered from start to finish.

🚀 Build A Software Business With Marcus — Click Here

📦 Get The Digital Product Dashboard Here — Build Your Business

Key Numbers From This Session:

Total spent on Manus AI: $3,200
Revenue attributed to Manus: $100,000+
Current monthly subscription: $400/month
Credits included at that tier: 85,000 per month

Section 1 — Why Manus AI Is Different: The Core Mindset Shift

Most people treat AI tools like a smarter search engine or a writing helper. Marcus treats Manus like a team of employees that work around the clock — simultaneously doing research, writing code, building files, and creating entire product ecosystems without needing to be babysat on every step.

The best way Marcus frames it: “If I could give you 10 employees for free this month — what would you have them do?” If you don’t have a fast answer to that question, you’re not yet thinking like a business owner. That’s the shift this whole training is about.

The key difference from other AI tools: ChatGPT and Claude answer questions and generate text. Manus executes multi-step tasks autonomously and delivers finished files — txt documents, PHP code, HTML pages, JSON files, entire folder structures — exactly the way you specify them.

Marcus spent over $3,200 on Manus across less than a year and conservatively attributes $100,000+ in revenue to that investment. His current plan is $400/month, which he says does the work of multiple outsourcers who would collectively cost $400–$1,000+ per week.

The right way to think about the cost: The question is never “Is this tool expensive?” The question is always “What is the output worth?” If $400/month in Manus spend generates $5,000–$20,000 in revenue, then the real cost is NOT using it. Marcus calls this the opportunity cost — the money you don’t make because you didn’t have the tool doing the work.

That said, he also warns strongly against frontloading tools before you’ve made any money. His rule: use PayPal buttons until you have sales, then add tools as revenue justifies them. He used standalone PayPal through 2011, making good money the whole time.

Section 2 — The 3-Platform Workflow Marcus Uses Every Day

Marcus doesn’t use just one tool. He has a defined workflow where each platform plays a specific role:

Step 1 — ChatGPT (Speed and Structure)
Used at the start of any project to generate fast structured lists, brainstorm frameworks, and organize raw research. ChatGPT is quicker for this than Manus because it doesn’t need to spin up agents or browse the web. You’re just getting a big structured list fast.

Step 2 — Manus AI (Execution and Bulk Building)
Takes the ChatGPT output and builds from it. Creates dozens or hundreds of coordinated files — text reports, JSON prompts, HTML pages, index files, code snippets — all in one run. This is the heavy lifting that would take a human team days or weeks.

Step 3 — You (Director and Architect)
Your job is not to do the work. Your job is to design the system, validate the idea, point the tools at the right targets, and connect the output to a monetization path. Like Henry Ford — he didn’t build the assembly line by standing in it. He designed it and placed the right pieces in the right spots.

Note on Claude: Marcus uses Claude specifically for building front-end HTML/PHP tools — the file viewers, popup interfaces, and index pages that turn a folder of text files into a usable software product. He says Claude is particularly good at clean, functional code for this kind of task.

Section 3 — Manus Superpower #1: Deep Bulk Research

No other AI tool does research at the scale Manus does. It doesn’t just answer a question — it goes out to the web, pulls structured data from multiple sources, processes it, and delivers organized reports and files. This is the equivalent of having a research team working in parallel.

What this looks like in practice:

Blog income research: Prompt Manus to find hundreds of blogs in a niche using income reports as the signal. It surfaces which blogs are making money, what niches they’re in, and how they monetize. This would take a human researcher days.
Affiliate program intelligence: “Find affiliate programs in [niche]. Extract payout amount, cookie duration, approval process, landing page URLs, and compile a report.” You get a complete affiliate landscape for any niche in one run.
Specific niche discovery: Instead of asking for “health” as a niche, ask for “100 specific profitable blog niches with data behind them — like losing weight after 40, not just weight loss.” The specificity is what makes it usable.
Competitor breakdown: Find the top blogs or sites in any niche, analyze what’s working, pull their traffic sources, content angles, and monetization methods into a structured report.
Keyword and problem research: “Find 100 problems homeowners face with pest infestations — warning signs, emotional triggers, what they search for, what they fear most.” This kind of empathy research used to take weeks of forum mining.
Social media analysis: Marcus has used Manus to analyze Pinterest — what niches are working, what pin styles get clicks, what topics are trending for income-generating content.
Bulk statistics: Pull data-backed numbers and citations across any topic to add authority to content, sales pages, or reports.

The BuiltWith comparison: Companies charge $500–$5,000/month to access organized data about websites. That data isn’t secret — it’s technically findable. What they’re selling is data structured and made useful. That’s exactly what Manus lets you do — take information that’s technically available and structure it into something people will pay for.

Section 4 — Manus Superpower #2: Bulk File and Asset Creation

This is where Manus becomes something completely different from other AI tools. It doesn’t write you one file — it produces hundreds of coordinated files in a single session, named exactly how you specify, formatted exactly how you need, ready to upload to a server and use immediately.

The live example from this session:

Marcus started with a ChatGPT list of 100 “special sauce” business frameworks — the structural advantage behind companies like Amazon, McDonald’s, Google, Netflix, and Zappos. He pasted that into Manus and gave it these instructions:

Create a full text report for each framework — what it is, how it works, what industries it applies to, which other businesses use a similar model
Create a JSON prompt file for each framework — a reusable prompt that applies that framework’s logic to any niche or business idea someone provides
Create an individual HTML page for each framework
Create one master index HTML file with clickable links to all frameworks
Name files by concept, not by company (e.g. “many-skus.txt” instead of “amazon.txt”) to avoid trademark issues

Result: 300+ files, completely formatted, with working HTML index. Time taken: 11 minutes. Credits used: approximately 271 — which at $400/month for 85,000 credits works out to roughly $1.28 total cost. His fastest outsourcer without AI would take 1–2 weeks for the same output. With AI tools, that outsourcer might take 2 days. Manus did it in 11 minutes.

The book framework project: Marcus took 250 non-fiction bestsellers, had ChatGPT extract why each one sold based on title, positioning, and promise, then sent that to Manus with instructions to write a full text report on each book’s success formula, create a reusable prompt for each that applies the same formula to any new book title, and build a clickable HTML index of all 250 with descriptions. The result is a ready-made product for authors — built in a single Manus session.

Important file strategy: Always request a master HTML index file — not a text index. A text index is just a list. An HTML index is clickable, browsable, and can be uploaded to a server and used as a live tool immediately.

Section 5 — Real Case Study: PersonalityPrompts.com

This is an actual business Marcus built with Manus, used as the primary live example throughout the training.

The origin story: Marcus built the initial version on a 3-hour drive to a concert in Tampa. He woke up early, started working, and had a functioning product ready to take to market within 5 hours. The business has made significant money since and he hasn’t done a major push on it since October of that year.

The core concept: Everyone is familiar with personality types — the 16 personality frameworks, DISC profiles, etc. Marcus took that further: what if you could distill the thinking style of successful entrepreneurs and business builders into AI prompts? Not “think like Elon Musk” (potential trademark issues), but what are the actual mental frameworks and decision-making patterns that made these people successful? Boil those into prompts anyone can use with any AI tool.

The technical setup — simpler than you think:

WordPress site on standard web hosting
A content folder on the server with organized subfolders (e.g. /content/copywriting/, /content/sales/)
Each subfolder contains .txt files — one per prompt or framework
A single PHP index file (built with AI in minutes) scans the folder and displays every .txt file as a clickable item that opens an inline popup with the content
Adding new content = drop a new .txt file in the folder. The index auto-discovers and lists it. Zero additional coding required.

The sales letter was also written by Manus — and it outperformed some of Marcus’s live sales webinars in conversion rate. He said that outcome surprised even him.

Why this model works: The product is low-overhead (text files on a server), infinitely expandable (drop in new files any time), and solves a real problem — people have AI tools but don’t know how to get strategic, specific results from them. The recurring revenue compounds. Marcus says 7pm every night is his favorite time of day because that’s when his recurring billing hits.

Section 6 — The Billion Dollar Pattern Engine: Built Live In This Session

Marcus built a brand new product concept from scratch during this training. Here is the complete step-by-step process:

Step 1 — ChatGPT: Generate the raw framework list
Prompt: “McDonald’s, Google, Amazon, and other major companies each had one specific thing that made them work in the beginning. Think about Amazon’s many SKUs, McDonald’s fast consistent systems, Google’s PageRank algorithm. Find 100 companies and list their secret sauce like this.” ChatGPT returns a numbered list of 100 companies with their structural advantage clearly named and described.

Step 2 — ChatGPT: Extract the entrepreneur value
Follow-up prompt: “How and why would understanding these frameworks be helpful to entrepreneurs and business people?” This gets you the value proposition of the product — why someone would pay for it.

Step 3 — ChatGPT: Name the product
Prompt: “What would be a great name for a tool that helps entrepreneurs run their idea through these frameworks to find if it has real leverage behind it?” Marcus chose: “Billion Dollar Pattern Engine” — because it says exactly what it does without hype or vagueness.

Step 4 — Manus: Bulk build the entire product
Paste the full 100-framework list into Manus with detailed instructions: create a text report per framework, a JSON prompt file per framework that applies that framework to any business idea the user provides, individual HTML pages for each, and one master HTML index. Name files by concept, not by company name.

Step 5 — Claude: Build the front-end file viewer
Ask Claude to build an HTML/PHP file system that auto-scans the folder and displays all .txt files as clickable items with popup content windows. No database needed. No manual updates ever. Just drop files in and they appear.

Step 6 — Upload and launch
Everything goes into one folder on web hosting. Gate it with PayPal or AMember for $7–$47/month and you have a live product.

The prompt quality lesson: Marcus ran one of the prompts in ChatGPT to test it and it didn’t work right on the first pass. He iterated the prompt language until it applied the framework directly to whatever business idea the user submitted. He spent time here because the prompt quality IS the product. The files are just the delivery mechanism.

Section 7 — Content Multiplier: From One Idea To a Full Marketing Ecosystem

Once the core product files exist, Marcus immediately runs a second Manus session to turn that same content into a full marketing and content library. One good idea multiplies into weeks of content without additional creative work.

Social media batch production: He sent the 100 framework list to Manus with these instructions: “Turn each item into a social media post. Lead with a hook that stops the scroll. Then describe the method and who used it. Include a ‘learn more in comments’ call to action. Create 3–5 comment templates that add more value, with the last comment linking to personalityprompts.com. Format as one big HTML page with a checkbox I can tick when each one is posted.”

The output: 100 ready-to-post social media posts organized on a single page with checkboxes, so a VA can work through the list and mark each one done. Every post leads people back to the product.

Short video script production: Same content, new Manus run: “Turn each framework into a 30-second video script. Lead with a hook. Describe how the method works. Make it engaging and educational.” The output is 100 video scripts covering 100 different business frameworks — all pointing back to the product at the end.

The content math from one Manus session: 100 frameworks → 100 text reports → 100 social posts → 100 video scripts → 100 email angles → 100 infographic concepts. All from one afternoon. All pointing to one product. All built before spending a dollar on ads.

The “Product Is The Marketing” strategy: When your product is built around frameworks, patterns, and insights, your marketing content IS your product. Every post about “here’s what made McDonald’s dominate” is both free value and a direct sample of what members get inside. No hard sell needed — the content demonstrates the value automatically.

Section 8 — Employee Thinking vs. Investor Thinking

One of the most important mindset concepts from the entire session. Most people using AI tools are still thinking like employees — they want to complete one task faster. Profitable AI use requires thinking like an investor — building systems that produce assets.

Employee: Write a single blog post. Investor: Build an entire niche content business with 100 posts.
Employee: Fix one sentence. Investor: Create 100 complete assets in one Manus run.
Employee: Generate a tweet. Investor: Build a complete product and market it simultaneously.
Employee: One slow task at a time. Investor: Run three business builds in parallel Manus sessions.
Employee: Trading time for output. Investor: Multiplying yourself — Manus runs while you sleep.
Employee: “Is this tool expensive?” Investor: “What is the output worth and what am I losing by not having it?”

The Henry Ford analogy: Henry Ford didn’t hire workers and hope an assembly line appeared. He designed the system, then placed people exactly where they needed to be. You are the architect. Manus is the assembly line. Your job is to design what gets built and why — not to stand in the line yourself.

Section 9 — Finding The “Magic Sauce”: What Actually Makes Businesses Work

The deepest concept in this session. Marcus argues that most entrepreneurs are chasing surface ideas — the what — when all the money is in understanding the structural why. The companies that dominate do so because they found one specific advantage and built everything around it.

McDonald’s: Never about the food. It was about a fast, systematic process that produced consistent output every time. Speed + consistency + replicability = the franchise model that changed everything.
Google: Hundreds of search engines existed before Google. What made them different wasn’t that they searched — it was that their algorithm ranked results by relevance better than anyone else. One structural advantage = a trillion-dollar company.
Amazon: Started with books because books have millions of unique SKUs. More unique products = more unique search terms = more organic traffic = dominance at scale. The product category was chosen strategically, not randomly.
How To Win Friends and Influence People: Probably not the scientifically best book on social skills. But the title promised exactly what the reader wanted in plain language. The positioning was the magic sauce, not the content.
Simple Sites Big Profits: Marcus’s first major digital product. A friend suggested the name. That name — with the value proposition built directly into the title — made the product work. A bad name would have killed it regardless of the content quality.

What Manus does with this insight: It extracts these structural patterns from hundreds of businesses, packages them into frameworks, and helps you apply those frameworks to validate or sharpen your own idea. That’s the real value of the Billion Dollar Pattern Engine — not 100 business facts, but 100 proven structural lenses you can apply to any niche or concept.

Section 10 — How The Tech Actually Works: Front End and Back End Explained Simply

Marcus demystifies the technical side completely. His core point: every internet business is just data on screens. Once you understand that, building software products stops being intimidating.

Front end: What users see — the webpage, the tool interface, the popup content viewer. Built with HTML and PHP. Claude or ChatGPT can build it in minutes from a plain English description.

Back end: What makes it work — the folders of .txt files, .json prompt files, the PHP script that scans the folder, the web hosting that serves it all. Manus creates these files. Standard web hosting ($5–$15/month) serves them.

The connection: A single PHP index file scans a folder for .txt files and displays each one as a clickable item that opens a popup with the content. That is the entire software behind PersonalityPrompts.com at its core. Adding new content = drag and drop a new .txt file into the folder. The index auto-discovers it. No code changes ever needed.

Payment options:

Simplest: PayPal button that unlocks a page URL. No shopping cart needed to test the concept.
Marcus’s preferred: AMember — about $150 one-time, handles memberships, access levels, and recurring billing. He’s done millions in sales with it.
Start lean: Marcus used standalone PayPal buttons until 2011 — over a decade and millions in revenue. The idea and offer matter infinitely more than the tech stack.

Section 11 — Important Warnings and Reality Checks

Most people who try to make money online make nothing. Marcus states this clearly and repeatedly. These results come from 20+ years of iteration, failure, reinvestment, and genuine expertise. This is not a shortcut to wealth.
AI will validate bad ideas. ChatGPT tends to be encouraging. If you ask it whether your idea is good, it will almost always say yes. Real validation comes from the market — people paying real money.
Don’t frontload tools. Buying every tool before you’ve made a dollar is one of the most common and expensive beginner mistakes. Earn first, then add tools as revenue supports them.
Rights and usage matter. Don’t scrape, repackage, and resell data you don’t have rights to use commercially. Add significant transformation and original value. Be careful with celebrity names and trademarked brand names.
Manus still requires direction. It is a blank canvas. Without a clear business objective and specific instructions, you will produce busy work — files that don’t add up to anything sellable.
Focus beats volume every time. Marcus is upfront about his tendency toward ADD. His most profitable periods come when he locks into one specific thing he knows will work and ignores everything else until it’s done.

Section 12 — Step-By-Step Action Plan: What To Do After This Training

Step 1 — Pick one niche or audience, and only one.
One specific group with one specific problem. The more specific you are, the faster everything else works.

Step 2 — Find the structural advantage that would help that audience.
What frameworks, patterns, or research does your audience need? What does the market charge a lot for that you could deliver better or faster? Use ChatGPT: “What information would [audience] pay for that isn’t easy to find or organize?”

Step 3 — Use ChatGPT to build a fast structured list.
Generate 100 items — frameworks, niches, problems, book titles, hooks. Get it in a numbered list you can copy. This is your raw material for Manus.

Step 4 — Send that list to Manus with detailed file instructions.
Specify every file type you want: text reports, JSON prompts, HTML pages, master index. Specify naming conventions and what each file should contain. The more specific your instructions, the better the output.

Step 5 — Use Claude to build the front-end viewer.
Ask Claude: “Build a PHP + HTML file that scans a folder for .txt files and displays them as a clickable list. When clicked, show the content in an inline popup.” Upload both files to your hosting folder.

Step 6 — Run Manus again for social content and video scripts.
Same list, new session. “Turn each item into a social media post with a hook and call to action.” Then another run: “Turn each into a 30-second video script.” Your entire marketing pipeline is ready before you spend a dollar on ads.

Step 7 — Get your first sale with a simple PayPal button.
Don’t build an elaborate funnel before validating the concept. Put a PayPal button on a simple page. When real people pay real money, you’ve proven the concept. Then invest in building it further.