Built a Full SaaS Product With an AI Agent — Here's What Actually Happened
I'm a non-technical founder (well, semi-technical — I can read some basic code, but writing it from scratch isn't my thing).
A few weeks ago, I launched Belink, a free link-in-bio tool that competes with Linktree.
The entire thing was built using Manus, an AI agent that writes code, deploys it, and manages your project end-to-end.
I want to share what the experience was actually like — the good, the surprising, and the stuff nobody talks about.
What Belink Is
Belink is a link-in-bio platform.
Users can:
Sign up and claim a username
Customize their page with themes
Add links
Embed YouTube videos
Share their page anywhere
There's a free tier and a Pro plan ($5/month) with Stripe billing.
The stack:
Next.js (Pages Router)
PostgreSQL (Neon)
Drizzle ORM
tRPC
Tailwind CSS
Vercel deployment
I didn't write any of that. Manus did. 😁
How I Actually Work With Manus
It's not “type a prompt and get a finished app.”
It's more like having a senior developer on call 24/7 who works insanely fast.
My workflow looks like this:
I describe what I want in plain English
(“Add email verification to signup using Resend”)Manus reads the codebase, understands the architecture, writes the code, runs migrations, and pushes to GitHub
Vercel auto-deploys
I test it, find issues, and report them
Manus fixes things in minutes
The back-and-forth is where the magic happens.
I'm not writing specs or PRs. I'm literally chatting like I would with a coworker:
“The favicon should be circular on the profile page.”
“Block signups from disposable email domains.”
What Surprised Me👍
Speed
Features that would take freelancers days were getting shipped in 10–15 minutes.
In one afternoon, I added:
IP-based rate limiting
Disposable email blocking
Admin pagination
Dynamic OG images
Context Retention
Manus remembers the entire project.
I don't have to re-explain:
Database schema
File structure
Existing architecture
It already understands how everything connects.
It Handles the Boring Stuff
Database migrations.
Cache-busting query strings.
CORS issues.
Meta tags.
All the tedious engineering work that normally slows projects down just gets handled.
It Catches Things I Miss
I noticed spammers were still bypassing my email blocking system.
Manus analyzed the database and figured out they were abusing Indonesian free subdomains like:
It then blocked those TLD patterns automatically.
That kind of investigative debugging honestly surprised me.
The Honest Downsides
You Still Need Product Taste
Manus builds what you ask for.
If your product decisions are bad, you'll simply get a very well-built bad product.
The “what to build” part is still completely your responsibility.
Debugging Requires Good Communication
Sometimes a feature doesn't work as expected.
You need to explain problems clearly.
“It's broken” doesn't help.
“Users can still register more than 3 accounts from the same domain” gets results.
You Become the QA Team
Since you're not writing the code yourself, you need to test everything manually.
I constantly use the product myself to find edge cases.
The Numbers
Time from zero to launch: ~2 weeks
Total features shipped: 50+
Monthly cost: Manus subscription + Vercel/Neon free tiers
Lines of code I personally wrote: 0
Would I Recommend This Approach?
If you're a solo founder with strong product sense but limited engineering skills — absolutely.
It's the closest thing to having a technical co-founder without giving up equity.
If you're already a developer, it's still incredibly useful as a force multiplier.
Think of it like pair programming with someone who:
Never gets tired
Never complains
Types at 500 WPM
The key mindset shift:
You're not “prompting AI.”
You're managing a developer.
Give clear requirements.
Test the output.
Iterate fast.
That's it.
Belink is live at belink.to if you want to check it out.
Happy to answer questions about the workflow.
Replies