From zero to five live SaaS products in one week. Here is what I learned, what broke, and what I would do differently.
The Challenge
I wanted to test: can one developer, armed with Claude and Next.js, ship real products in a week?
The answer: yes, but with caveats.
The 5 Products
- AccessiScan (fixmyweb.dev) - WCAG accessibility scanner, 201 checks
- CaptureAPI (captureapi.dev) - Screenshot + PDF generation API
- CompliPilot (complipilot.dev) - EU AI Act compliance scanner
- ChurnGuard (paymentrescue.dev) - Failed payment recovery
- DocuMint (parseflow.dev) - PDF to JSON parsing API
All built with Next.js, TypeScript, Tailwind, deployed on Vercel.
What Worked
- AI for boilerplate code (auth, API routes, UI components)
- Vercel for instant deployment
- Upstash Redis for rate limiting and usage tracking
- Stripe for payments (surprisingly easy to integrate)
What Did Not Work
- Trying to make everything perfect before shipping
- Building features nobody asked for
- Spending too long on design before validating demand
The Numbers (Honest)
| Product | Pages | Build Time | Revenue |
|---|---|---|---|
| AccessiScan | 40+ | 8h | $0 |
| CaptureAPI | 40+ | 6h | $0 |
| CompliPilot | 40+ | 10h | $0 |
| ChurnGuard | 45+ | 12h | $0 |
| DocuMint | 40+ | 8h | $0 |
Yes, zero revenue so far. Building is the easy part. Finding customers is the hard part.
Lessons Learned
- Ship fast, iterate based on feedback
- AI accelerates coding 3-5x but you still need to understand what you are building
- The European Accessibility Act creates real demand for accessibility tools
- Payment recovery is a real problem - 30 percent of SaaS revenue is lost to failed payments
- Distribution matters more than product quality
What I Would Do Differently
- Start with one product, not five
- Find 10 potential customers BEFORE building
- Use cold email outreach from day one
- Focus on SEO content from the start
All products are live with free tiers. Try them out and let me know what you think!
Building in public at toolkitonline.vip
Top comments (1)
5 in 7 days is a great velocity flex, and it surfaces the most interesting shift AI enables: the bottleneck moved from "can I build it" to "which of these is worth building." When shipping is cheap, the scarce resource becomes distribution and validation, not engineering. The risk of the 5-in-7 approach is ending up with 5 orphans that each got zero users - velocity without a distribution plan is just faster way to make things nobody sees.
So the meta-question I'd ask after a sprint like this: of the 5, which one got a real user, and what would it take to go deep on that one instead of wide on five? That's actually a workflow I optimize for with Moonshift (a multi-agent pipeline that ships a prompt to a deployed SaaS) - make each build cheap and fully deployed (~$3 flat, real auth/billing/deploy, first run free) so you can run lots of validated shots on goal, then double down on whichever one shows signal. Build wide to find the winner, go deep to make it real. Impressive output. Did any of the 5 get traction, and are you planning to focus or keep spinning up new ones? The portfolio-vs-focus decision is the fun part after a week like that.