I Built 3 SaaS Products With Vibe Coding — Revenue, Costs, Lessons

On a Thursday morning in March, I opened the Stripe dashboard on my phone while waiting for a flat white at a cafe on Oranienstrasse. The number at the top of the screen — total monthly recurring revenue across all products — had crossed a threshold I had been watching for weeks. Not a life-changing number. Not a quit-your-day-job number for most people. But for a solo founder who could not write a line of code eighteen months ago, it was the kind of number that made my hand shake slightly as I held the phone, and the barista had to say "flat white" twice before I heard her.

4,340 euros per month. Across three products. Built entirely with vibe coding.

I am going to open the books on all three — what they cost to build, what they cost to run, what they earn, and what they taught me. This is the post I would have wanted to read before I started: not the highlight reel, not the "I made $50K in a weekend" bait, but the real, line-by-line picture of what it looks like to build a software business alone, with no technical background, using AI as your only engineering team.

Product 1: Soulin Social — The Content Multiplier

Soulin Social was the first thing I built. It started as a tool for myself — I was spending fifteen hours a week creating and distributing content across five platforms, and every hour on content was an hour not spent on the work that actually mattered. I needed a machine that could take one raw thought and turn it into platform-ready posts that sounded like me, not like a chatbot.

What it does: Takes a single input — usually 50-150 words of raw thinking — and generates 35+ platform-specific posts in my trained voice. LinkedIn, X, X threads, Instagram, Substack intros, email snippets. Each formatted for its platform, each calibrated to my specific writing patterns.

Time to build: First working version took a weekend. Production quality took two months of nightly iteration.

The stack: Node.js, Claude API for generation, Supabase for user data and content storage, Vercel for hosting, Stripe for payments.

Monthly operating cost: $47 (Supabase Pro $25, Vercel $8, Claude API usage ~$14 depending on volume).

Revenue (March 2026): 2,900 euros MRR. 94 paying users at a mix of 29 and 49 euro tiers. 93 free-tier users. Monthly churn: 8.2%.

Time invested per week (current): About 10 hours. Bug fixes, feature requests, support, and content about the product. Some weeks more when things break. Some weeks less when they do not.

The cost to build: Zero in direct development costs — just my existing $20/month Claude subscription. But the time cost was enormous. Two months of 3-5 hour nightly sessions, seven days a week. If I valued my time at even 30 euros per hour, that is roughly 12,000 euros in time investment. I do not regret it. I just want to name it, because "I built this for free" is a lie when your nights disappear.

The biggest lesson: I launched with eight platform formats and the quality was thin across all of them. When I cut to five and did each one well, both quality and retention improved immediately. Focus is not a philosophy. It is a business decision with a measurable P&L impact.

Product 2: KINS Sales Agent — The AI Closer

KINS is the wellness hotel brand I built before I ever touched code. The sales process was drowning me — 2-3 hours daily answering the same fifteen questions from potential guests. "What protocols do you offer?" "How long should I stay?" "What does it cost?" Factual, repetitive, exactly the kind of work an AI should handle.

What it does: An AI chatbot embedded on the KINS website that handles initial sales inquiries. It responds with accurate information about healing protocols, room availability, pricing, and logistics. When a conversation signals booking intent — specific dates, payment questions, explicit interest — it escalates to me via Telegram with a full conversation summary.

Time to build: One week to first version. Three weeks total to production quality.

The stack: Node.js, Claude API for conversation, Supabase for conversation storage and knowledge base, custom chat widget, Telegram for escalation.

Monthly operating cost: $32 (API usage varies with conversation volume — some months as low as $18, the highest was $41 during peak booking season).

Revenue impact: The KINS Sales Agent does not generate revenue directly — it does not have its own subscription model. Its value is indirect but massive. It handles approximately 70% of initial inquiries without my involvement. My response time to booking-ready guests dropped from 24-48 hours to under 2 hours. The conversion rate from inquiry to booking improved by roughly 35% because guests get immediate, accurate, warm responses instead of waiting for me to find time between building software and running a hotel.

In concrete terms: the agent contributed to approximately 2,800 euros in additional bookings per month that I estimate I would have lost to slow response times. I cannot prove that number precisely — there is no control group for your own life — but the before/after booking data is clear.

Time invested per week (current): About 2 hours. Mostly updating the knowledge base when protocols or pricing change. The agent itself runs quietly. When it breaks, it breaks loudly — silence from the chat widget means something is wrong — so monitoring is simple.

The biggest lesson: The first version was too eager to escalate. I was getting fifteen Telegram notifications a day, most for conversations that were just browsing. I tightened the escalation rules — checking for multiple signals instead of single ones — and now I get three to five notifications a day, almost all of which convert. Precision in automation matters more than coverage.

Product 3: Soulin SEO Agent — The Autonomous Monitor

The SEO agent is different from the other two. It is not user-facing. It is infrastructure — an autonomous system that monitors search rankings, tracks keyword performance, analyzes competitors, and sends me Telegram alerts when something needs my attention. It runs on an EC2 instance via PM2, 24 hours a day, whether I am awake or not.

What it does: Crawls search results for my target keywords daily. Tracks ranking changes over time. Monitors competitor content for relevant terms. Analyzes on-page SEO for my published posts and suggests improvements. Generates weekly reports. Alerts me via Telegram when rankings drop significantly or when a new opportunity emerges.

Time to build: Two weeks to first functional version. Ongoing iteration since — this one never feels "done" because SEO is a moving target.

The stack: Node.js, Claude API for analysis, Supabase for data storage, PM2 for process management, EC2 for hosting, Telegram for alerts.

Monthly operating cost: $38 (EC2 instance $15, Claude API for analysis ~$18, Supabase shared with other products, Telegram free).

Revenue impact: Like the KINS agent, the SEO agent does not charge subscription fees. Its value is in the organic traffic it helps me earn and protect. Since deploying it, organic traffic to soulin.co has grown by roughly 180% over four months. Some of that is content volume — I have been publishing more. But the agent's keyword research and on-page optimization suggestions have directly influenced which topics I write about and how I structure posts.

In estimated revenue terms: organic traffic accounts for approximately 40% of new Soulin Social signups. At current conversion rates, that is roughly 640 euros per month in MRR attributable to organic search — traffic that the SEO agent helps me earn and maintain.

Time invested per week (current): About 3 hours. Reviewing alerts, acting on suggestions, occasionally debugging the agent when a scraping pattern changes and the parser breaks.

The biggest lesson: Autonomous agents need guardrails. The first version alerted me for every ranking change — including fluctuations of a single position that meant nothing. I was getting forty Telegram messages a day. I built threshold logic: only alert for drops of three or more positions, or for new top-10 rankings. The noise dropped by 90%. The signal stayed.

The Full Picture

Here is the cost table across all three products:

	Soulin Social	KINS Sales Agent	Soulin SEO Agent	Total
Monthly revenue/impact	2,900 EUR (direct)	~2,800 EUR (indirect)	~640 EUR (indirect)	~6,340 EUR
Monthly operating cost	$47	$32	$38	$117
Build time to MVP	1 weekend	1 week	2 weeks	—
Build time to production	2 months	3 weeks	Ongoing	—
Weekly maintenance	~10 hrs	~2 hrs	~3 hrs	~15 hrs
Development tools cost	$20/mo Claude + $20/mo Cursor	(shared)	(shared)	$40/mo

Total monthly operating cost including development tools: roughly $157.

Soulin members get the full essay library, private group chat, the Soulin OS e-book, and every tool — all for $10/mo. Join Soulin →

Full essay library · Private group chat · Soulin OS e-book · Every tool · $10/mo

Total monthly revenue and revenue-impact: roughly 6,340 euros.

These are not venture-scale numbers. A VC would look at this table and close their laptop. But here is what the table does not show: I own 100% of everything. I have no investors, no board, no one to report to. My total capital investment was zero euros — just time and existing subscriptions. My monthly profit margin is over 97%. And the 15 hours per week I spend maintaining these products is less than the 20+ hours per week I used to spend on the tasks they replaced.

The Hidden Costs Nobody Talks About

The operating costs are clean and low. The hidden costs are messier.

Time cost of learning. The first three months, I was not building efficiently. I was learning how to prompt, how to debug, how to think in systems. Hours disappeared into rabbit holes that produced nothing shippable. If I tracked all the hours from "first Claude conversation" to "first dollar of revenue," the number would be somewhere around 400 hours. At 30 euros per hour, that is 12,000 euros of invisible investment. It does not appear on any balance sheet but it is real.

Mental health cost. I do not know how to quantify this, but I want to name it. The 2am debugging sessions. The week the payment integration failed silently and I ate the loss. The month I drafted a shutdown email for Soulin Social because the support load was crushing me and the revenue was not enough to justify the damage to my sleep. Solo building is lonely in a way that is different from other kinds of loneliness — you are the only person who understands the system, the only person who will notice if it breaks at 3am, the only person responsible when something goes wrong. That weight has a cost.

Opportunity cost. Every hour I spent building software was an hour I did not spend on KINS operations, on writing, on the healing work that is the actual foundation of everything I do. The products generate revenue, but the question I ask myself regularly is whether that revenue justifies what I gave up to earn it. Most months the answer is yes. Some months it is less clear.

I list these not to discourage — I would make the same choices again — but because the "I built this for $20 a month" narrative erases the real price. The subscription fee is the smallest cost. The time, the sleep, the loneliness, the opportunities you decline because you are debugging instead — those are the actual expenses.

The Compound Effect

What the numbers do not show is velocity. Each product made the next one faster and cheaper to build.

Soulin Social took two months to reach production quality. KINS Sales Agent took three weeks. The SEO Agent took two weeks to first functional version. The improvement is not because the products got simpler — the SEO agent is architecturally the most complex of the three. The improvement is because I got better at the meta-skill: describing systems to an AI, anticipating failure modes, testing incrementally, debugging efficiently.

The same compound effect applies to maintenance. Early on, every bug was a crisis. Now most bugs fall into patterns I recognize. "The API response format changed" — I have fixed that before. "The database connection timed out" — I know the checklist. "The deployment failed on a new dependency" — I have a process for that. The knowledge accumulates and the panic recedes.

This compounding is the real argument for vibe coding as a career strategy, not a one-time experiment. The first product is expensive in time and stress. The second is cheaper. The third is cheaper still. By the fourth — which I am planning now — I expect the path from idea to production to take days, not months.

What These Numbers Actually Mean

I am not building toward an exit. I am not trying to reach $10K MRR or $100K ARR or whatever milestone the SaaS Twitter crowd has decided is the threshold of legitimacy.

I am building toward enough.

Enough to cover my Berlin rent — 780 euros. Enough to cover my operating costs. Enough to fund the next product without borrowing. Enough to not take meetings I do not want to take, work on projects I do not believe in, or trade my time for someone else's priorities.

4,340 euros in direct MRR — with another roughly 2,000 in indirect revenue impact — is enough. Not comfortable. Not luxurious. Enough, in the way that enough means you are not afraid of the first of the month.

The real lesson from these three products is not about vibe coding. It is about the math of independence. You do not need millions. You do not need hockey-stick growth. You need a handful of products that solve real problems, cost almost nothing to run, and generate steady revenue that covers a life you have intentionally kept small.

My life is small on purpose. One apartment, one laptop, one person. The smallness is not a limitation — it is the strategy. Small costs mean small revenue is sufficient. Sufficient revenue means freedom. Freedom means I build what I want, when I want, for the people I want to serve.

That is the number that matters. Not 4,340. The number that matters is the distance between what you earn and what you need — and three vibe-coded SaaS products turned that distance, for me, from a chasm into a step.

What would "enough" look like for you — and how many of the things standing between you and that number are actually necessary?

I write about freedom, healing, and building alone. The full archive is at soulin.co.