Building Boring Tools That Rank: 4 Weeks, 11 Tools, and a Plot Twist with ChatGPT

Published: (December 11, 2025 at 06:43 AM EST)
4 min read
Source: Dev.to

Source: Dev.to

The Premise

I’m a system engineer from Romania working a 9‑5. My side income comes from client work, not products. I wanted to test if a solo developer could build something that grows while sleeping, powered entirely by SEO.

No cold outreach. No paid ads. No Product Hunt launches. Just ranking pages that solve problems people search for every day.

The tools themselves are nothing revolutionary: merge PDFs, compress images, resize photos, extract colors. The kind of utilities millions of people need daily. The experiment was whether I could make them rank.

Week 1: Starting the Engine

Shipped the first two tools

  • PDF Merge
  • PDF Compress

Both process files entirely in the browser. No uploads to servers. No tracking. Everything happens locally.

Set up the basics: Next.js, Firebase analytics, proper metadata, JSON‑LD schemas. Submitted to Google Search Console and waited.

Lesson learned – New domains have zero authority. Google doesn’t care about you yet.

Week 2: Shipping Fast

Tools added (now 5 total)

  • Split PDF
  • Image Resize
  • Image to PDF

SEO fixes

  • Meta titles over 65 characters (truncated in search results)
  • Missing canonical URLs
  • No internal linking between tools

Converted the site to a PWA and started getting the first installs.

Key insight – Tool pages without content don’t rank. Added guides, FAQs, and keyword‑rich paragraphs to every page.

Started backlink outreach to tool review sites, PDF comparison blogs, and tech writers.

Biggest insight – Free plugins equal free backlinks. WordPress plugins, Chrome extensions, npm packages—these directories send links. iLovePDF and SmallPDF have been doing this for years.

Week 3: First Signs of Life

Milestones

  • First 100+‑visitor day
  • Brand searches appearing on Bing (people typing “quicktools.one” directly)
  • Multiple tools crawled and indexed within 48 hours of launch

New tools

  • Color Picker
  • Delete PDF Pages
  • Compress Image

Added a “smart suggestions” feature: after using one tool, the page suggests related tools. Tiny change, huge impact on retention and internal linking.

Traffic milestone – Pacing at ~3,000 monthly users (organic + returning).

Submitted to Bing Webmaster Tools and Yandex. Crawl frequency improved across all search engines. Multi‑search‑engine discovery can accelerate indexing.

Week 4: The Plot Twist

Traffic source breakdown

SourceVisitorsShare
ChatGPT18071%
Google5622%
Direct155%

ChatGPT sent 3× more traffic than Google. An AI chatbot became the biggest traffic source without any optimization for it.

Weekly stats

  • 675 unique visitors (+51 % week over week)
  • 841 visits
  • 3,080 page views
  • 18 % bounce rate
  • 3 min 56 sec average session

New tools

  • Crop Image
  • PDF Extract
  • PDF to Image

Total: 11 tools live in 4 weeks.

Why Is ChatGPT Recommending My Site?

  1. The tools actually work – No sign‑ups, no email gates, no dark patterns. Just tools that do what they say.
  2. Privacy positioning stands out – Most PDF tools upload files to servers. Processing everything locally in the browser is a real differentiator.
  3. Clean UX signals quality – High engagement metrics (82 % of visitors interact, nearly 4‑minute sessions) probably matter to AI systems crawling user feedback.
  4. Presence on developer platforms – Packages on npm, extensions on Chrome Web Store and Edge Add‑ons signal legitimacy to AI systems.

What Actually Worked

  • Internal linking is the fastest SEO win. Every tool now links to 3‑5 related tools, creating crawl paths, topic clusters, and longer user sessions.
  • Static HTML links matter more than hidden JS. If a link isn’t visible in the HTML at load time, crawlers ignore it.
  • UX leads to SEO faster than expected. Behavior signals (time on page, interaction) boost search‑engine trust.
  • Blog posts support keywords and backlinks. Comparison articles started pulling impressions after 2‑3 weeks.
  • Multiple discovery channels reduce risk. Traffic from ChatGPT, Bing, npm, extension stores, and direct visits creates resilience; relying only on Google is dangerous.

The Numbers After 4 Weeks

MetricValue
Tools live11
Monthly visitors (projected)~3,000
Bounce rate18 %
Avg. session duration3 min 56 sec
Top traffic sourceChatGPT (71 %)
Backlinks~20‑50 (varies by tool)
Revenue$0 (not monetized yet)

What I’d Do Differently

  • Track everything from day one. Missed logging directory submissions, so I can’t measure which ones worked.
  • Start backlink outreach earlier. Domain authority is the bottleneck; the sooner you build links, the faster pages rank.
  • Don’t underestimate AI discovery. Building genuinely good products might be the optimal strategy for both SEO and AI recommendations.

What’s Next

  • Video‑to‑GIF converter (high‑volume keyword)
  • More blog content for long‑tail keywords
  • Exploring why ChatGPT recommends certain tools
  • Continuing the npm/extension backlink strategy
  • Eventually: monetization with lightweight ads

Try the Tools

Everything is free and open:

Follow Along

I post daily updates on X: @eduardalbu

If you’re building something small and trying to rank it, I’d love to hear what’s working for you.

Tags: #buildinpublic #seo #indiehackers #webdev #nextjs

Back to Blog

Related posts

Read more »