From renovation sites to React deployments — a story of building software with Claude and ChatGPT
My name is Kostiantyn. I've spent 20+ years in construction and interior design. I run a design studio, manage renovation projects, calculate estimates. And every single time a new project started, I opened the same spreadsheet, copied columns, fixed formulas, made mistakes, recalculated.
One day I did the math: creating one estimate from scratch takes me 3-4 hours. Finding current material prices — another hour. Calculating how many bags of plaster I need — calculator, notebook, manufacturer's website, comparisons. Every project — starting from zero.
I'm not a programmer. Never took an IT course. Don't know React, don't know SQL, never heard of Supabase. But I had something else — 20 years of understanding exactly what a person standing on a construction site with a tape measure actually needs.
So I built a SaaS platform using AI.
The Market That "Doesn't Exist"
Construction estimates in Ukraine — and frankly, in most of Eastern Europe — exist in two extremes with nothing in between.
On one end: professional desktop software for government-regulated estimates. They cost $300+, run only on Windows, and serve project institutes and large developers. A regular contractor doing apartment renovations doesn't need them.
On the other end: CRM systems for construction businesses. Full ERPs with IP telephony, warehouse management, accounting, document flow. Starting at $50/month. One such competitor gathered 65 users in 3 years — the market speaks for itself.
In the middle — the contractor. The person who actually does the renovation. They coordinate a crew, purchase materials, communicate with clients, control the budget. There are tens of thousands of them. And 90% still work in spreadsheets, notebooks, or messengers.
Why? Because no tool existed for them. Government estimate software — too complex and expensive. A $50/month CRM — too many features they don't need, and too few they actually do: quickly calculate, show the client a professional document, protect your margin.
What I Built
LIHTAR is a cloud platform for construction estimates. Mobile-first — it works from your phone on the construction site. Here's what it does:
Dual pricing. The contractor sees the real (working) price. The client sees their price — with a margin the contractor sets. The client never sees the actual cost. Sounds basic, but no competitor in the Ukrainian market has this. And for a contractor, this is the #1 pain point — how to not reveal your profit to the client.
AI material calculator. 9 types of construction work: floor screed, plastering, wall prep, ceiling prep, tiling, drywall structures, ceilings, masonry, tile laying. Enter the area, layer thickness, choose the manufacturer — get a complete material list with consumption rates based on technical specifications. Not "approximately 5 bags" but a precise calculation accounting for layers, coefficients, and consumables.
AI import. Have an old estimate in any format — text, photo, scan? AI recognizes the items and creates a structured document.
Gantt chart linked to estimate items. Not just a calendar of "when's the measurement, when's the installation" but a project management tool showing which subcontractor enters the site when.
ClientView — the client receives a link, sees their estimate (client price, no margin visible), work schedule, and can communicate via chat. No registration required.
Market price directory — 783 items with automatic updates. The contractor always knows the current price per m² for plastering or tile laying.
Branded PDF. The crew sends the client a professional branded document with company logo, not a photo of a notebook.
All of this costs ~$7/month. Or free for one project.
How I Built It Without Programmers
Honestly: through Claude and ChatGPT.
I described what I wanted in plain language. "I need a table with two price columns — working and client — where the difference is the margin, but the client only sees their price." AI generated React code. I downloaded the files and replaced them in the project folder, pushed to GitHub, checked the result, described what was wrong, AI fixed it.
Stack: React 19, Supabase (database and auth), Vercel (hosting and deployment), GitHub. Everything cloud-based, everything free or cheap to start.
Sounds simple? No. It was hell. Every second AI response had a bug. Every third component broke another one. I rewrote single files 50+ times. GanttTimeline.js — a file I'm terrified of because every change can break everything.
But in 50 days I built a product with 30+ features that works, has real users, and ranks #1 on Google Ukraine for "estimate software."
50 days. Solo. No investment. Through AI.
Will AI replace programmers? No. I need a developer and I'm looking for one. But AI let me validate the idea, build an MVP, and enter the market before I hire a team. It's like building the first floor of a house yourself to see if anyone wants to live there — then hiring a crew for the rest.
Competition in a Niche Where Everyone Sleeps
The market for cloud-based commercial renovation estimates is essentially empty.
There's a CRM at $50/month with IP telephony, warehouse, and accounting. A full business tool for companies. But a contractor doesn't need a warehouse and accounting module. They need to calculate an estimate and send it to the client.
There's an online service that gathered 65 users in 3 years. UX with horizontal scrolling, not mobile-adapted, no AI, no dual pricing, no material calculator.
There are dozens of calculators on building material store websites — primitive tools designed to sell bags of cement.
And there's LIHTAR — which in its first month acquired as many users as one competitor did in three years.
It's not because I'm a genius. It's because the niche is empty, the product solves a real pain, and the price is right.
What's Next
I'm finalizing the paid subscription system. Running Google Ads with a ~9% CTR — unusually high for B2B SaaS. Applied for the Startup EDGE grant (up to €40,000). Preparing partnerships with building material manufacturers — they want to be in the calculator where contractors actually calculate their products.
Goal for 2027: 1,000 paying users. Sounds modest for a tech startup. But in a niche where the entire market still counts in notebooks — it's a revolution.
Construction is one of the largest sectors of the economy. And one of the least digitized. Tens of thousands of contractors spend hours daily on what could be done in minutes. Not because they're lazy — because the tool didn't exist.
Now it does.
LIHTAR — lihtar.app
Author — Kostiantyn Yakubiv, founder of LIHTAR, 20+ years in construction and interior design
Tags: #startup #saas #ai #nocode #construction #proptech #builtwithAI
Top comments (0)