The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
What I Check Before Starting Any New App Project
Before writing code, I always take time to understand the product clearly. A good app is not only about clean screens or working features — it needs the right foundation, a smooth user flow, and a clear reason for users to keep coming back.
The first thing I check is the core goal of the app. What problem are we solving? Who is the user? What action should they take first? If this part is unclear, even a well-built app can feel confusing.
Next, I review the main user journey. I look at how users will sign up, navigate, complete important actions, and return later. Small friction points can hurt the experience, so I try to identify them early before development becomes expensive.
I also check the technical requirements: authentication, database structure, APIs, payments, notifications, admin panels, third-party integrations, and future scalability. This helps prevent rushed decisions that may create problems later.
Another important part is understanding the MVP scope. Many projects try to build too much at once. I prefer to define the must-have features first, launch a strong initial version, then improve based on real user feedback.
Finally, I check communication and expectations. Clear milestones, fast feedback, and honest updates are just as important as the code itself. When the client and developer are aligned from the beginning, the final product becomes much stronger.
For me, starting a project the right way means reducing risk before development begins. A successful app is not built by rushing into code — it is built by understanding the product, the users, and the path to launch.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
My full design cycle for an AI startup 💻✨
One of my recent favorite projects was a full design cycle for Telephone AI —
brand identity, website, industry pages, and the platform itself.
AI integrated into every step of the process.
Case study is still coming 👀
But working on that project reminded me of what I actually love most
about early-stage startups: being in on it early.
Researching a new market, figuring out who the real user is,
what they need, what they've never had before —
and laying the right foundation from day one
instead of fixing it later.
Best practices from MVP. Not retrofitted after.
Here's what that process looks like for me now:
↳ Claude Projects — all the context in one place:
call transcripts, competitor research, personas,
feature priorities for MVP and what-can-wait
↳ Claude Code — structure, interactions,
clickable logic tested before Figma even opens.
Design system with tokens built from the start.
↳ Figma MCP — moved there for customization
and a proper clickable prototype the client can actually walk through
From idea to something real to click —
without weeks of "give us time to think."
#ProductDesign #StartupDesign #AIWorkflow #UXDesign #Figma #ClaudeCode
Goated!
Kickoff, the companion that helps you start.
We don't struggle to do tasks so much as we struggle to begin them. That first step is the hardest part, especially with ADHD, burnout, or plain overwhelm. Kickoff takes whatever you're stuck on, breaks it into one tiny, doable first move, and a friendly character talks you through it, one step at a time. No streaks, no countdowns, no guilt. It rewards the bravest part of any task: starting.
How I built it:
I leaned mostly on Figma Make and designing directly on the Figma canvas. I started by building a small design system with an iterative approach, which gave me a solid foundation to carry through the whole project. From there I moved constantly between the canvas and Make, editing, refining, and expanding features as I went. I explored wiring up a Gemini API key for live step breakdowns and chat, and built a hardcoded fallback for when no API is connected, so the app always works. When I was nearly out of Make credits, I switched to Cursor using the Figma MCP to build out a few more features and make direct edits in code.
It started as a design system and grew into a working companion. Next up: turning it into a desktop plugin you build, keep, and can call at any moment.
Figma community link: https://www.figma.com/community/file/1649523053611199686/kickoff
Live website link: https://hem-relume-49421705.figma.site
Social post link: https://x.com/kpollyyk/status/2067862108755136916?s=20
I Take AI Slop and Turn It Into Product
At SanganakHQ, I've done this at least half a dozen times now. Same story every time.
Founder builds with Claude, Codex, Bolt, Replit, Lovable, v0. And it works. Real users, real payments, real data.
But under the hood? 50K, 100K+ lines of AI-generated code glued together. Functions duplicated dozens of times. Auth mixed into UI. Hardcoded secrets. No tests. Fix one bug, three more appear.
It works the way a house of cards works.
Then they message me: "We're scaling. Can you make this production-ready?"
So I audit every line. Separate what's solid (usually 20%) from what needs to be rearchitected. Build a day-by-day migration plan with exact pricing. Move it to a real stack: Next.js, Supabase, Vercel, proper auth, proper RLS, proper CI/CD.
The irony: the AI features are usually the cleanest part. The mess is everything around it. Business logic, payment flows, membership tiers. The stuff AI generated fast but never thought through.
Smart founders, solid ideas, AI got them to a working product. But "working" and "production-grade" are two very different things.
If your app runs on AI-generated code and you're about to scale, get it audited before you find out the hard way.
Trending
Claude
Claude has entered the design space. How are you using Claude Design?
Contra University
Learn from expert creatives how to earn more using next-gen AI tools.
MagicPath
The canvas is infinite, and exploration is becoming the workflow. How are you using MagicPath?
creativeaiflow
Creative AI workflows are evolving. What tools do you use, and what are their strengths and weaknesses?
freelancerlife
Freelancer life is wins, pivots, and everything in between. What’s yours right now?