I built 85 tools based on a fear I had myself
Two months ago I shipped Kastavo (kastavo.com) — a free tools portal (PDF, image, dev, calculators) that runs 100% in the browser. Zero backend, zero server processing.
The idea came from my own fear: downloading random "free PDF compressor" tools and worrying about filling my PC with malware — especially since I use my machine for programming. If that bothered me, a developer who's supposed to know better, it probably bothers non-technical users even more.
The zero-backend decision wasn't just a technical choice, it was the entire pitch: your file never leaves your device. No upload, no server-side processing, no "we delete your files after 24h" trust-me policy — because there's no server to delete anything from in the first place.
What validation actually looked like for me (no interviews, no landing page test):
- Searched where the pain was already being talked about (Reddit threads about sketchy PDF tools, malware complaints)
- Checked what alternatives existed — most competitors either require upload or aren't transparent about server-side processing
- Built it because I would use it myself daily
Where it's at 2 months in:
- 85 tools live, 3 languages (EN/PT/ES)
- Organic traffic growing (SEO + backlinks, zero paid spend)
- Just did a security pass — CSP, secure cookies, went from a B to a B+ on Mozilla's Observatory
Next up: a UX refresh (currently a pretty plain layout, redesigning with a proper visual identity) and I'm planning to keep shipping new tools regularly — a few categories I'm still deciding between right now.
Curious if anyone here has built something similarly "utility-first" — no accounts, no funnel, just tools people use once and leave. How do you think about growth for something like that?
85 tools gives you enough data to compare categories. Which signals are you using to decide where another tool is worth building versus doubling down on an existing cluster?
Google Analytics + Search Console, nothing fancier yet — at this stage I don't think I have enough volume to justify a more sophisticated stack.
The signals I actually look at:
Organic search queries hitting a tool that doesn't exist yet — Search Console shows what people are searching for on my domain (or adjacent terms) that I'm not currently serving. If the same non-existent-tool query keeps showing impressions, that's a build signal.
Time-on-page + bounce on existing tools within a category — if a tool in a cluster gets used but people bounce right after (no return visits, no related-tool clicks), that tells me the cluster itself might not be the problem, but something adjacent is missing.
Direct URL guesses in Analytics — people typing URLs like /tools/qr-code-generator before I've built it (shows up as a 404 in GA events). That's about the strongest signal I've found: someone assumed it existed.
Honestly still more art than science at 85 tools — I don't think I have enough data yet to do this rigorously. Curious if you've found a threshold (tool count, traffic volume) where this starts working like a real methodology instead of an educated guess?
That threshold question is actually worth discussing properly rather than squeezing it into the thread. Would email be easier? What’s the best address for you?
Email works, sure. kastavo@outlook.com — happy to go deeper on this.
Thanks! I’ve just sent it over.
Looking forward to hearing your thoughts whenever you have a chance.
The pitch is strong and it has one structural weakness: the whole promise is something a visitor cannot see. Every tool site claims it is private and an ordinary user has no way to check. The good news is yours is the rare claim that can actually be demonstrated. Turn off your wifi and the tool still works. That is one line of instruction on the page, it takes ten seconds, and no server based competitor can copy it. I would put that above the fold instead of explaining the architecture, because zero backend is a fact about you while works offline is a fact the person verifies for themselves.
That's a really sharp point — I hadn't thought about it that way. You're right that "zero backend" is a claim about me, but "works offline" is something the visitor can verify in ten seconds without trusting a word I say.
I'm going to test that instruction above the fold on the next update. Thanks for the reframe — it's a much stronger pitch than what I had.