Sign InSign Up

I Built a SaaS in 30 Days. Here’s Exactly What Happened.

AI Summary6 min read

TL;DR

I built StackTrace, a developer tool that summarizes production logs, in 30 days. Starting from zero, I reached 312 signups, 41 paying users, and $1,287 MRR by focusing on user feedback, launching transparently, and charging early.

Key Takeaways

  • •Focus on solving one painful problem simply rather than overbuilding with unnecessary features.
  • •Charge from day one to validate real user interest, not just curiosity.
  • •Transparency and honest communication drive more trust and growth than polished marketing.
  • •Build based on actual user requests rather than assumptions about what they want.
  • •Keep the tech stack simple to ship faster and avoid premature complexity.

Tags

buildinpublicsaaswebdevstartup

Thirty days ago I had:

  • No landing page
  • No logo
  • No users
  • No idea if anyone cared

Today I have:

  • 312 signups
  • 41 paying users
  • $1,287 MRR
  • A product I almost quit three times

This is the honest breakdown. No hype. No fake growth curve. Just what actually happened.

The Idea

I built a simple developer tool called StackTrace.

It turns messy production logs into readable incident summaries using structured parsing and AI tagging.

The problem was simple. When something breaks at 2 AM, nobody wants to scroll through 4,000 lines of logs.

You want:

  • What failed
  • Where it failed
  • Why it failed
  • What likely caused it

That’s it.

The Tech Stack

I kept it boring on purpose.

Frontend

  • Next.js 14
  • TailwindCSS
  • Vercel

Backend

  • Node.js
  • Express
  • PostgreSQL
  • Redis for rate limiting
  • OpenAI API for log summarization

Auth & Billing

  • Clerk for authentication
  • Stripe for subscriptions

Infrastructure

  • Railway for backend
  • Supabase for database
  • Cloudflare for DNS and caching

No Kubernetes. No microservices. No clever architecture.

Just ship.

Week 1: Build Fast, Build Ugly

Goal: Get a working prototype in 7 days.

What I shipped:

  • Log upload
  • Basic parsing
  • AI summary
  • Copy to clipboard button

What it looked like:

  • One page
  • No dashboard
  • No branding
  • Zero polish

I sent it to 14 developer friends.

Result:

  • 9 tried it
  • 4 said “this is actually useful”
  • 2 asked if it was paid

That was the first signal.

Week 2: Add What Users Actually Asked For

Instead of adding features I wanted, I only built what users requested.

Top requests:

  1. API endpoint instead of manual upload
  2. Slack integration
  3. Save previous reports
  4. Redact sensitive data automatically

So I built exactly those.

Metrics at end of week 2

  • 87 signups
  • 11 active weekly users
  • 0 revenue

Still validating.

Week 3: Launch on Dev.to

I wrote a transparent post about debugging production faster.

No marketing language. Just story plus screenshots.

It hit the front page for 12 hours.

Traffic spike

  • 6,421 visits in 24 hours
  • 201 new signups
  • 19 trial activations

That was the moment it felt real.

Pricing Experiment

I tested three models:

  1. $9 flat monthly
  2. Usage based
  3. Free tier plus $19 Pro

Flat monthly confused people.

Usage based scared people.

Free plus Pro converted best.

Current pricing:

  • Free: 20 log analyses per month
  • Pro: $19 per month unlimited
  • Team: $49 per month shared workspace

Revenue So Far

Day 30:

  • 312 total users
  • 41 paid
  • $1,287 MRR
  • 6 churned

Churn reason breakdown:

  • 3 said they rarely hit incidents
  • 2 found internal tooling sufficient
  • 1 just experimenting

Churn taught me more than signups.

The Biggest Mistakes

1. I Overbuilt Early

I wasted two days building a fancy analytics dashboard no one used.

Users just wanted summaries.

2. I Almost Added Microservices

At one point I convinced myself I needed a worker service, event queue, and separate AI processing pipeline.

Reality: I had 15 users.

Premature complexity is ego disguised as architecture.

3. I Avoided Charging Too Long

I waited until week 3 to turn on payments because I felt awkward.

That delayed real validation.

If someone will not pay, they are not validating your product. They are validating curiosity.

What Surprised Me

People Want Boring Tools

The most common feedback:

“This is simple but saves time.”

Developers do not need revolutionary platforms. They need friction removed.

Transparency Drives Growth

The Dev.to post brought more trust than ads ever could.

Developers can smell fake growth posts instantly.

Be honest and specific.

What I Would Do Differently

  • Charge from day 1
  • Build landing page before polishing UI
  • Talk to 20 users before writing serious code
  • Keep scope half as big

Current Challenges

  • Preventing AI hallucinated root causes
  • Reducing OpenAI API costs
  • Handling large log files efficiently
  • Deciding whether to build self hosting option

The next milestone is $3k MRR.

If I hit it, I’ll document the path.

If I fail, I’ll document that too.

Final Thoughts

Building in public is uncomfortable.

You expose numbers.

You expose mistakes.

You expose doubt.

But it forces clarity.

Shipping a real product in 30 days taught me more than 2 years of reading about startups.

If you are thinking about building something, do it small.

Solve one painful problem.

Charge early.

Talk to users constantly.

And document the journey.

Developers respect honesty more than polish.

Visit Website