CalcHive

CalcHive › Blog › How CalcHive Came to Be

How CalcHive Came to Be

From a YouTube rabbit hole on passive income to 36 live calculators in an afternoon. The honest version.

It started with a YouTube video

Like a lot of side projects, this one started by accident. I was watching YouTube videos on creating passive income and monetizing a website with Google AdSense, and I caught the bug. The idea seemed almost too simple: build something useful, drive traffic with SEO, let Google's ads do the rest.

The vision came fast. I wanted a one-stop shop for all things conversion, calculation, and estimation. Not just one calculator, not just five, but a whole library. Cooking conversions, mortgage payments, time math, BMI, paint coverage. Anything you might Google in a hurry on your phone, in one place, free and fast.

Attempt one: the manual era

This was before I was using Cowork. I was using the Claude chat web app. I started a project, sketched out the vision, and got to work the only way I knew how at the time.

That meant manually creating the files. Claude would write code, I would copy it into a TextEdit file, save it as index.html, build out the folder structure by hand, and drag everything to Netlify Drop to see if it worked. Every time Claude suggested an update, I had to go back into the files and embed the changes by hand. Every fix was a copy-paste-save-redeploy cycle.

It was workable, honestly. Screenshots helped a lot. I could show Claude exactly what was on screen and get pretty good guidance back. But without a programming background, the whole thing was clunky at best. I was the bottleneck. Every change had to pass through me, and I was just shuttling text from one window to another.

I got to an MVP. Five calculators on a single page, live at a random Netlify URL. But I had a quiet worry I never quite shook: I was not confident the formulas were all correct. I could not really audit the math without being able to read the code I was deploying. Was this conversion accurate to the third decimal? Was the percent change formula doing the right thing with negative starting values? I was guessing.

So I did what most people do with side projects: I put it on the back burner.

The detour

I spent the next stretch learning other things. AI automations. Prompt engineering. n8n, Make.com, Zapier patterns. Voice agents. I was casting a wide net, trying every tool I could find. My browser had a tab for every AI service that launched.

That sprawl is what eventually brought me back to Claude. Anthropic kept shipping new models and tools, and at some point I started thinking I could consolidate. Why was I jumping between five chat interfaces and three automation platforms when one of them was getting noticeably better and adding the features the others had?

So I installed Claude Desktop and started learning Cowork.

The afternoon that changed everything

Cowork was the unlock. With the API connectors, the Chrome extension, and the ability for Claude to actually edit files on my computer (instead of me copy-pasting between windows), the whole equation flipped. Claude was not just writing code anymore. It was building the site.

I sat down one Friday morning and told Claude to pull the original calculator project off my computer. By bedtime, I had:

That same night, I added two more high-value calculators (Social Security retirement estimator and rental property cash flow), then went back later and built out 27 more. By the end of the build, CalcHive had 36 live calculators across 7 categories: Finance, Cooking, Measurement, Time, Math, Health, and DIY.

And, critically, every single formula had been audited against known reference values. The mortgage calculator agreed with bank amortization tables down to the cent. BMI matched the standard formula in both imperial and metric. The Navy body-fat formula returned values inside the expected range for test inputs. Length, weight, temperature, and volume conversions used official conversion factors (28.3495 grams per ounce, 0.3048 meters per foot, the full Imperial vs US fluid ounce distinction, all of it). The credit card payoff calculator was fixed mid-audit when we found it was over-estimating interest by 3 percent due to an assumption that the final payment would be the full payment instead of the actual amortized remainder.

This is the part that mattered most to me. The original concern that stalled me out the first time around (am I really sure the math is right) was solved by being able to actually inspect the code, run reference calculations against it, and fix issues as I found them. With Cowork, I could ask Claude to verify the math, see the results, and have the fix applied to all 36 files in one operation. The math audit alone would have been days of manual work in the old setup. It took 15 minutes.

What CalcHive actually is

The vision did not change much from the original sketch. CalcHive is the place you go when you need a quick answer and you do not want to fight a recipe blog or a finance site for it. The person I picture every time I build a new tool is a school teacher with four minutes between classes, looking up a measurement conversion for the science demo this afternoon. Time-pressed, mobile, smart but not technical. No accounts, no popups, no signup walls.

That persona shaped every decision: big tap targets, fast page loads, no above-the-fold ad walls, result-first display, explanations underneath for people who want them. The brand color (honey amber) and the hexagonal mark came from the "hive" idea: a busy collection of practical tools, all working together. It also gave us calchive.org, which turned out to be available after we walked away from the original name's brand collision.

What I learned from doing this twice

The biggest lesson is not really about Claude specifically. It is about where AI sits in the workflow.

In version one (Claude chat web app, manual file management, copy-paste between windows), I was the bottleneck. Every change required me to translate from one medium to another. AI was a smart assistant, but the workflow around it was 1995-era. Most of my time was spent NOT building.

In version two (Claude Desktop with Cowork, API connectors, file editing), the bottleneck moved. I am still the one making decisions and giving direction, but I am no longer the courier between rooms. Claude reads files directly, writes files directly, runs scripts directly, edits in place. The clunkiness disappears.

That feels like the real shift. AI tools have been "good enough" for content for a while, but the friction was always in the wiring around them. Cowork is the first setup I have used where the wiring matched the AI's actual capability.

If you are someone who has tried to build something with AI and walked away frustrated because the workflow was the problem (not the AI), this is worth knowing.

What is next for CalcHive

The site will stay 100 percent free. Always. The whole positioning is built on "no accounts, no paywalls, just answers" and that promise is the moat. I will monetize with AdSense and, eventually, with affiliate or lead-gen partnerships on the high-intent finance calculators (mortgage referrals, financial advisor matching). But there will never be a "CalcHive Pro" subscription. If a more advanced tool ever needs to exist, it will be a separate product under a different brand.

For now, the focus is on adding more calculators (a Take-Home Pay calculator, Retirement Projection, Calorie Burn with proper MET tables, Time Zone Converter, Macro Calculator, and more), writing the blog content that helps people use them well, and getting indexed by Google so the SEO can compound.

If you want to follow along

I write a newsletter on Beehiiv about how I am using AI tools to build, automate, and run small businesses. If the story above resonated (the part about going from sprawl to consolidation, or the part about how a workflow shift changes what is possible), you would probably get something out of it.

Subscribe to the newsletter

Or, if you just want to use the calculators that started all this: browse CalcHive. New tools land every week.

← Back to all posts