I left my job, knew nothing about coding, and shipped a SaaS in 2 weeks. Here is the short story.

In October 2025, I went back to a company I worked for before. I believed in the product. But slowly, I felt pushed into a corner to just deliver outputs.

One Thursday in early February, I was asked to build a backlog of 20 items by Monday, just so engineering had something on their table to work on. Something inside me snapped. I tried to explain we needed time to think about outcomes, not just fill a backlog. But the answer was: "We cannot waste engineering time in discussions."

My confidence eroded. Not in the product, but in myself. We were choosing speed over clarity. So, I left.

The itch I couldn't scratch
I always was curious to build things. I had the product instinct from years as a PM, but I could not code. Usually, I looked for a technical co-founder. This time, I decided to do it alone.

I took a week to learn Claude Code and Cursor. The idea came from a pain I had for years as a Product Manager. People always asked "how do I do this?", and I would manually take screenshots, paste them in a doc, and write text. It was draining. Tools like Scribe or Tango exist, but they are heavy, built for massive standard operating procedures. I wanted something simple, fast, and lightweight.

So I started building Quiqlog.com. Not because I wanted to make it a big thing, but simply out of curiosity to learn how to build things with AI tools.

The first version was a mess
I prompted Claude Code with vague instructions and got vague results. I had no idea how to manage staging VS production environments. And the Chrome Web Store submission and review porocess was a total headache.

Eventually, I made a decision that felt terrible but was actually right. I threw the whole thing away and started from scratch.

Building it right
The second time, I was disciplined and specific. Instead of vague prompts, I wrote proper PRDs - as a Product Manager I know how to build good ones. I split the work into small chunks and started to prompt Claude for each small piece of work separately. In total, I gave Claude over 600 prompts and did 50+ deployments to staging before pushing to production.
I stopped saying "Claude fix this" and started saying "Claude, investigate this and list the reasons why it happens." I learned that the quality of Claude Code output is almost entirely about the quality of how you instruct it to do something, what details/inputs you give it.

The two days that almost broke me
The hardest moment was figuring out how my Chrome extension talks to the right environment. For two full days, things worked in staging but broke in production. I was going in circles. I didn't even know the right questions to ask.
Finally, I stopped describing symptoms. I asked Claude to explain the full picture of how environments work in the code. In minutes, I had the answer.
I felt surprised in a good way, like: "Wow, so the problem was simple, but I was not asking the right way about it."

The Stack
For anyone curious, here is how I built it:

• Web App & API: Next.js
• Database & Auth: Supabase (the backbone of the product)
• Payments: Polar
• Hosting: Vercel (with auto-deploy from GitHub)
• Extension: Vanilla JavaScript (no frameworks)
• Landing Page: I tried Framer first, but it was too frustrating to navigate all the tiny details. I switched to Lovable. It is still a work in progress, but it works for now.

Where I am now
Quiqlog.com is live. Real users can record workflows and share guides in minutes. I built something from nothing, alone. That feels significant.
My next step is to build one more app purely solo: a lightweight CRM for individual travel experience hosts. Right now, they manage travelers and vendors using messy Google Docs, spreadsheets, and messengers. I want to ship something just to learn more.

Seeking advice
But after that, I have a big decision to make, and I would love your advice. I am at a crossroads. Do I keep going down this solo building path? Or do I take this new perspective of building things as a Product Manager and look for a new Product Manager role in a team with a true, outcome-driven culture?

If you have been in my shoes - building your own things versus joining a great product team - what is your advice? What should I do next? Let me know in the comments.