Hello ProductHunt community,

I am currently developing a service that combines Agentic E2E Testing with Automated Triage (Fail Reporting), and I would love to hear your honest feedback.

Demo/Showcase Introducing Jaebau

The Problem

Hey everyone! I'm getting ready to launch Reality Try, a Virtual Try-On app for Shopify.

The crazy part? I don't have a traditional backend engineering background. I essentially "vibe coded" the entire backend logic and AI API integration using GitHub Copilot, and set it up to scale on GCP.

It was a wild process of bridging the gap between my product vision and the actual execution using AI as my "lead developer."

Has anyone else here successfully built and launched a complex app relying mostly on Copilot or Cursor? What were the biggest roadblocks you hit when moving from prototype to production? Would love to hear your experiences!

If you've heard this halfway through a demo call, that's the problem.

Too many sales calls happen just to explain the basics.

Sales teams repeat the same walkthrough before getting to real questions.

The AI learning platform that actually builds your thinking, not your dependence.

You've done it. We all have.

You're stuck on something a concept, a problem, a piece of writing so you open ChatGPT, type your question, and copy the answer. Problem solved. Task done. Move on.

This weekend holiday, I built something called Nudge.

As developers, we often talk about productivity but our systems are usually cluttered:

Too many tabs.

Too many background apps.

Hey everyone! I think writing GPU code is going to be something that's going to become ubiquitous even though today it's a tough and time consuming process. Claude Code and Cursor agents have gotten great at writing GPU kernels once they know what to target. What's missing is the tooling to get them there.

We're building that tooling at https://ncompass.tech. Our agent analyzes the profile and gives back a list of bottlenecks and ways to solve them. Once that's known, you can feed that to Claude Code / Cursor and they will implement the code.

In 2026, this is a very fair question:
If I can just use free ChatGPT/Grok/Gemini to upload my CSV and analyze sales, why do I need Okiela?
*The short answer:
ChatGPT is a brilliant general-purpose tool.
Okiela is a specialist for e-commerce profit analytics.
-------------
A few concrete examples from real Shopify/Lazada/TikTok Shop data:
Okiela already understands the structure of ecommerce exports. You don t have to explain what Lineitem price , Subtotal , or Created At mean every single time. You just upload, wait a few seconds, and you get a proper profit dashboard.

Shopify exports repeat Subtotal/Shipping/Tax on every line of an order. If you simply sum Subtotal , most AI will overcount revenue 2 5x. Okiela hard-codes Shopify logic so these fields are only counted once per order.

Southeast Asian currencies are messy: 1.500.000 is one point five million VND, not 1.5 . Okiela has dedicated logic to correctly parse VND/THB/MYR/PHP formats and avoid subtle but deadly mistakes.

When COGS is missing, Okiela will say N/A upload COGS data , instead of silently treating cost as zero and making you think profit equals revenue.

And most importantly:
Okiela is not just a chat . It gives you:
1. A shareable dashboard,
2. KPI cards, channel charts, SKU tables,
3. Export to PDF/Excel (Pro),
4. AI chat that runs on top of that structured dataset.
So in my mind:
-Use ChatGPT when you need a super smart general AI.
- Use Okiela when you want fast, reliable, shareable profit insights for your ecommerce store.
You can try the free tier (3 analyses/month, no card) at https://okiela.io and see the difference in the first 30 seconds.
------------------
#buildinpublic #analytics #data #shopify #ecommerce #saas #solofounder #Okiela

Hey guys,

I wanted to share my project - I hope this forum topic will be a correct one.
I built TMDD - an open source CLI that keeps a version-controlled threat model (YAML format) inside your repo and generates security-aware prompts for AI coding agents.

So what is threat model? It is a simple document where you write down what you re building, how someone could abuse or break it, and how you ll stop that from happening. You usually also include data flows diagram inside of it. Some argue that it's the most efficient method of detecting security issues in early phases of development.

When you vibe code with AI, it usually focuses on does it work? , not on Can someone exploit this? .

CodeMasterIP updated again this week, adding many more learning lessons, achievements, and profile customization! Todo en www.codemasterip.com

I recently published my Android app Daily Reminder & Task Manager and I m looking for people who enjoy trying new productivity apps and sharing genuine feedback.

What I m looking for:
Usability feedback (what feels confusing or slow)
Feature suggestions
Bug reports
Real experiences after using it for a few days

This is purely for improving the app no incentives, just honest opinions.