DEV Community: 蔡旭

How Urlcomp reverse-engineers an AI startup from one URL

蔡旭 — Thu, 10 Sep 2026 04:14:45 +0000

Most “competitor research” for an AI product is a pile of tabs: the pricing page, a Similarweb screenshot, a ChatGPT thread, and a guess about whether you should build a clone.

Urlcomp is built for a narrower job. You paste a public product URL at https://www.urlcomp.com and get a 10-layer competitive report in minutes — public pricing, clone difficulty, growth signals, and a founder build / no-go. Every field is tagged Found, Inferred, or Not Public. It is not a traffic dashboard, not a keyword rank tracker, and not a chat window.

Pricing is on the site: guests get 3 analyses per device per month; registered Explorer gets 3 per day (preview layers, not the full L1–L10). Single Report is $9 for one complete teardown (3-day access). Builder is $29/month (200 analyses). Founder is $79/month (300 analyses + Intelligence Hub). Start free, no credit card: https://www.urlcomp.com/pricing

Here is a real sample, not a mock UI:

https://www.urlcomp.com/report/notion.so

What the 10 layers are

L1 Product Snapshot — hard facts (positioning, public pricing, stack, traffic sources).
L2 Why Users Pay — why people pay; the moat.
L3 Business Engine — how money moves; unit economics as estimates.
L4 Build Similar Product — MVP path, clone difficulty, cost.
L5 Opportunity Intelligence — gaps and underserved segments.
L6 Growth Signals — GitHub, Product Hunt, hiring, momentum.
L7 Startup Timeline & Intelligence Feed — recent pricing / feature / positioning changes.
L8 Competitor Snapshot — direct and indirect rivals.
L9 Startup Radar — keep watching after the first teardown.
L10 Founder Intelligence — should you build, is it too late.

The tags matter more than the layer names

Found = verified from a public source (Notion’s public price table is Found).
Inferred = an estimate with reasoning shown (a monthly-revenue field on the sample is Inferred — treat it as a hypothesis, not audited financials).
Not Public = no reliable public source; the cell stays empty instead of a fake number.

If you only remember one habit: trust Found, challenge Inferred, do not fill in Not Public yourself.

green check = Found, gray circle = Inferred, red “Not public” = Not Public.
A smaller hook if you are not ready for a full report: Clone Check is a free preview of clone difficulty and public pricing.

https://www.urlcomp.com/tools/clone-check

Privacy, in one paragraph: live user reports stay in the account and out of search. The Notion / Cursor / ChatGPT pages are official samples, deliberately indexable.

What it is not

Similarweb estimates traffic and channels. Urlcomp reverse-engineers one product.
Semrush is search and ads. Crunchbase is company and funding files. Urlcomp is neither.
RivalLens at getrivallens.com is a different page-monitoring product. Urlcomp was previously published at rivallens.net; the public name is now Urlcomp.

If you are an indie hacker staring at an AI product and asking “can I clone this, and should I?”, start with the Notion sample, then run Clone Check on a public URL you actually care about.

Site: https://www.urlcomp.com
Notion sample: https://www.urlcomp.com/report/notion.so
Cursor sample: https://www.urlcomp.com/report/cursor.com
Pricing: https://www.urlcomp.com/pricing

Previously at rivallens.net. Not RivalLens (getrivallens.com).

How Urlcomp reverse-engineers an AI startup from one URL

蔡旭 — Thu, 10 Sep 2026 04:10:26 +0000

I Reverse Engineered Lovable From a Single URL — Here's What I Found

蔡旭 — Thu, 27 Aug 2026 06:14:27 +0000

TL;DR

Lovable opportunity: 7.8/10, but market entry 3/10 with 7/10 saturation.
It's moving from "AI code gen" to build, collaborate, deploy, and run.
Don't build another general-purpose AI app builder.
Better wedges: vertical-specific, mobile/app-generation, enterprise (security, compliance, data residency).
I used one URL + Rivallens to track product moves, competitors, and founder decision.

Most startup research starts with dozens of tabs.

Product pages. Pricing. Competitors. Social media. News.

But I wanted to try something different:

How much can you actually learn about an AI startup from a single URL?

So I ran Lovable through a startup intelligence analysis, starting with nothing more than its product URL.

I wasn't trying to write another product review.

I wanted to answer a more useful question:

If I were a founder, should I build in this market?

Here's what I found.

1. Lovable Is Becoming More Than an AI Code Generator

Lovable's capabilities now span:

Natural-language → full-stack app generation
Real-time preview and editing
Database and authentication integration
Deployment and hosting

The product is moving beyond "generate some code for me" toward "help me go from an idea to a working application."

The L9 product timeline also surfaced these as active product moves, positioning Lovable as an AI-powered web app builder for non-technical users.

That creates a bigger opportunity—but also a much harder competitive environment.

2. The Market Is Attractive. Entering Head-On Isn't.

The analysis gave Lovable's opportunity landscape a 7.8/10 Opportunity Score.

At first, that sounds like a great market to enter.

But another signal tells a different story.

Rivallens scored the broader market 3/10 for entry attractiveness, with 7/10 saturation and around 7/10 feature overlap.

Why?

Because Lovable isn't alone.

The competitive landscape includes:

Bubble · Webflow · Retool · Glide · Bolt.new · v0.dev · Replit

So the conclusion isn't:

This is a bad market.

It's:

There is opportunity—but don't compete head-on.

3. Where Is the White Space?

Three potential wedges stood out to me.

Vertical-specific AI builders

Build for healthcare, finance, education, or e-commerce.

The differentiation isn't just better AI generation. It's:

AI + industry workflows + integrations + templates.

Mobile app generation

As general-purpose AI web-app builders become more crowded, stronger mobile app generation could become an interesting wedge.

Enterprise AI builders

Enterprise customers care about requirements beyond AI generation:

Security
Compliance
SSO and audit logs
Data residency

These create room for a more specialized product.

There are also adjacent gaps around cost predictability, migration paths, and bridging AI-generated code with traditional development workflows.

4. The Growth Signals

The L9 startup radar showed 3 tracked changes, including a pricing change and a feature launch.

Pricing was confirmed at:

Free / $20 Pro / $50 Business

The 30-day timeline also surfaced product activity around agent/feature updates and team collaboration.

The direction is interesting.

Lovable appears to be expanding from:

Build something

toward:

Build, collaborate, deploy, and run something.

That makes the product more valuable to serious users—but raises the competitive bar for anyone trying to enter.

5. So, Should You Build a Lovable Competitor?

The build verdict was:

Do Not Enter / Wait

But the founder recommendation wasn't simply "stay away."

It was:

Proceed with differentiation.

Avoid head-on parity with Bubble, Webflow, Retool, Glide, and other established players.

The better strategy looks more like:

❌ Another general-purpose AI app builder

↓

✅ Vertical-specific AI builder

✅ Mobile/app-generation-focused builder

✅ Enterprise-focused AI builder

The opportunity isn't necessarily to beat Lovable.

It may be to build around the gaps that a company like Lovable creates.

The Bigger Insight

A normal startup analysis tells you:

What does this company do?

A more useful analysis asks:

What's changing?

Where is the market opening up?

Where is competition getting stronger?

Should I enter?

And those answers change over time.

That's why startup research shouldn't be a one-time report.

Why I Built Rivallens

I built Rivallens to make this process faster and repeatable.

Start with a product URL.

Then analyze its:

Market
Competitors
Opportunities
Growth signals

And keep tracking what changes.

For Lovable, that meant going from:

URL → Market Intelligence → Opportunity → Founder Decision

instead of manually opening dozens of tabs.

Analyze once. Monitor continuously. Build with confidence.

Try Rivallens → https://rivallens.net

What AI startup should I reverse engineer next?

Drop a URL in the comments.

I'll pick one.