I Had a €1k Budget and Built a SaaS That Repurposes AND Publishes Your Content
The honest build story behind ContentMill — from zero to a live SaaS with AI repurposing, direct publishing, Stripe billing, and a waitlist, all on a €1,000 budget.
Most "I built a SaaS" posts skip the parts where everything breaks. This one won't.
I started with a simple thesis: creators spend more time reformatting content than creating it. A blog post becomes a LinkedIn carousel, a Twitter thread, a newsletter intro, an Instagram caption — all manually. It's tedious, repetitive, and it doesn't scale.
So I set out to build a tool that doesn't just repurpose content — it publishes it too. Here's exactly how that went.
The Budget: €1,000
Not a metaphor. That's the literal credit card limit I started with.
Here's how it broke down:
- Vercel Pro: $20/mo — hosting, serverless functions, edge network
- Supabase: Free tier — Postgres database, auth, row-level security
- Claude API (Anthropic): ~$50-100/mo — the AI engine behind repurposing
- Stripe: 2.9% + 30¢ per transaction — payment processing
- Domain (contentmill.net): ~$12/year
- Resend: Free tier — transactional emails
Total monthly burn: roughly $70-120. The €1k budget gives me about 6-8 months of runway before the product needs to pay for itself.
The Tech Stack
I went with boring technology that ships fast:
- Next.js (App Router) — React framework with server components, API routes, and static generation in one package
- Supabase — Postgres + Auth + Row Level Security. No backend to build.
- Claude API — Anthropic's model handles the actual content repurposing. It's remarkably good at understanding platform conventions and tone
- Stripe — Checkout, subscriptions, webhooks. The ecosystem is unbeatable
- Tailwind CSS — Utility-first CSS. No design system needed, just ship
- Vercel — Git push to deploy. Zero DevOps
Why these choices? Because I'm one person with a credit card. I need managed services that scale from zero to thousands of users without me touching infrastructure. Every hour spent on DevOps is an hour not spent on the product.
Tired of manually rewriting content for every platform?
ContentMill turns one blog post into LinkedIn, Twitter/X, and newsletter content in 45 seconds.
The MVP (Week 1)
The first commit was the core repurposing engine. Paste in a blog post, get back platform-optimized versions for LinkedIn, Twitter/X, Instagram, YouTube, and newsletters.
That's it. No accounts, no billing, no dashboard. Just the transform.
The AI prompt engineering was the hard part. Each platform has different conventions:
- LinkedIn wants hooks, short paragraphs, and a question at the end
- Twitter/X needs threads with numbered posts, each under 280 characters
- Instagram requires emoji-rich, hashtag-heavy captions
- YouTube needs SEO-optimized descriptions with timestamps
- Newsletters need a personal, conversational opening
Getting Claude to reliably produce all five formats from a single input took more iteration than I expected. The trick was being extremely specific in the system prompt about each platform's conventions and giving concrete examples.
The Insight That Changed Everything
Here's what I learned talking to early users: repurposing content is only half the problem. The other half is actually posting it.
People would generate their LinkedIn post, copy it, open LinkedIn, paste it, format it, add hashtags, and schedule it. Then do the same for Twitter. And Instagram. And their newsletter.
The repurposing saved 10 minutes. The publishing still took 30.
So I built direct publishing. Connect your LinkedIn and Twitter/X accounts, and ContentMill posts for you. One click, multiple platforms. This was the feature that turned "nice to have" into "I need this."
What I Shipped (In Order)
Here's the rough timeline of features, compressed into about three weeks:
- Core repurposing engine — AI-powered content transformation
- Supabase auth + user accounts — Email/password, magic links
- Stripe billing — Free (5 repurposes), Pro ($29/mo), Business ($79/mo)
- Landing page — Conversion-focused, with live demo
- Direct publishing — LinkedIn and Twitter/X API integration
- Post scheduling — Queue posts for optimal timing
- Instagram publishing — Business API integration
- Custom brand voice — Paste writing samples, AI matches your tone
- Content intelligence — Learn from post performance over time
- Onboarding flow — Reduce time-to-first-repurpose
- Blog + SEO pages — Content marketing infrastructure
- Team workspaces — Collaborate on content (Business plan)
- Referral program — Viral loop built into outputs
- Waitlist + early access — Email capture with live counter
- Team tier ($49/mo) — Bridge the gap between Pro and Business
- Founder's Deal — 40% off for life for early annual subscribers
That's 16 major features in ~3 weeks. Shipping fast isn't about working 16-hour days — it's about choosing tools that eliminate busywork.
The Hard Parts Nobody Talks About
Stripe webhook reliability. Your checkout flow can be perfect and still fail if your webhook handler doesn't account for duplicate events, out-of-order delivery, or the customer closing their browser mid-checkout. I spent an embarrassing amount of time debugging a whitespace character in my Stripe secret key.
AI output consistency. Claude is incredibly capable, but "creative" means it sometimes surprises you. Getting reliable, platform-correct output every time required very specific prompts, output validation, and stripping markdown artifacts from the final output.
Social platform APIs. LinkedIn's API documentation is... aspirational. Twitter/X changes their API terms frequently. Instagram requires a Business account connected through Facebook. Each integration was a rabbit hole.
Free tier abuse. Five free repurposes sounds generous until someone creates 50 accounts. Rate limiting and fingerprinting are ongoing work.
The Numbers (Honest)
Since I'm writing this for Hacker News, here are real numbers:
- Total spend to date: ~€300 of the €1,000 budget
- Time to MVP: ~1 week
- Time to full product: ~3 weeks
- Lines of code: The codebase is a single Next.js app. No microservices. No separate backend. One repo.
- Infrastructure: Zero servers. Everything is serverless/managed.
- Deployment: Git push → Vercel auto-deploys in ~30 seconds
What I'd Do Differently
Start with publishing, not repurposing. The "aha" moment for users is seeing their content go live on LinkedIn with one click — not seeing the repurposed text. I should have built the publishing integration first and the repurposing second.
Charge from day one. I started with a generous free tier (originally unlimited, then 5 repurposes). The people who give you the best feedback are the ones paying you money. Free users give you feature requests. Paying users tell you what's broken.
Write the blog content earlier. SEO takes time to compound. I waited until week three to start writing blog posts. Should have been week one.
What I Learned About PMF
Product-market fit isn't a moment — it's a gradient. Here's how I'd describe the stages:
- "That's cool" — People see the demo and nod. (Week 1)
- "Can I try it?" — People want access. (Week 1-2)
- "I'll pay for this" — People enter credit card details. (Week 2-3)
- "I told my friend" — Organic referrals start. (TBD)
I'm somewhere between stage 2 and 3. The waitlist is growing. Early users are converting to paid. But the referral loop isn't spinning yet.
The honest answer: I don't know if this will work. But the economics are simple. At $29/mo, I need about 4-5 paying users to cover costs. At 50 users, I have real revenue. At 500, I have a business.
The Stack I'd Recommend
If you're building a SaaS in 2026 with minimal budget:
- Next.js + Vercel — You won't outgrow this for a long time. Server components are a genuine productivity unlock.
- Supabase — Postgres without the ops. Auth without the headaches. RLS means your API routes don't need authorization logic.
- Claude API — If your product has any AI component, Anthropic's models are the best combination of quality and developer experience.
- Stripe — Nothing else comes close for billing. The Checkout product alone saves weeks of work.
- Tailwind — Stop debating CSS architecture. Just ship.
Total cost to run this stack at low scale: $50-100/month. That's less than most people spend on coffee.
Try It
ContentMill is live at contentmill.net. Paste in any blog post and see it transformed for 6 platforms in seconds.
Free tier gives you 5 repurposes. Pro is $29/mo for unlimited repurposes and direct publishing.
If you're a creator, marketer, or founder who's tired of reformatting content — this is built for you.
Questions? I'll be in the comments.