Little Internal Apps With a Big Punch

What happens when building software becomes too easy??

For years, when I wanted a tool to handle a specific, annoying thing in my business, I had two options: find a SaaS that sort of did it and pay $40 a month forever; pay a freelancer to build it; or build it myself the hard way and lose a weekend (let’s be honest, a whole month). So mostly I just… didn’t. I lived with the annoying thing.

Today that calculus has completely changed. Between AI doing the heavy lifting on the code and Cloudflare letting me host damn near anything for free, I can now spin up a scrappy little internal app in an afternoon to scratch exactly one itch. So I’ve been doing that. A lot. Maybe too much, my wife would say too much.

Here’s the show-and-tell: the little internal apps quietly running my business right now. None of them are products. Most of them are held together with duct tape and good intentions. All of them punch way above their weight.

CRM & email marketing automation

This one started as a “wait… could I just build this?” moment. It’s a proof of concept that we could self-host our own bespoke email marketing tool (list management, automations, the works) instead of renting one. It runs on a database and media storage sitting on Cloudflare, and it already does the core job: catch a lead, tag it, and run it through a sequence. Is it going to dethrone the big marketing-automation platforms tomorrow? No. But it proved the thing I wanted to prove: the expensive monthly tools aren’t magic, and for a lot of small businesses, “good enough and ours” beats “fancy and rented.”

The great thing here is that I steadily build features out one at a time, and for those things I can’t get to? They open a GitHub issue to be spec’d by a GitHub bot that automatically specs issues with a specific label.

Positioning Advisor

I bang the drum constantly about a laser-focused positioning statement, the “I help ___ with ___” thing. The Positioning Advisor is the tool version of that whole conversation. It walks you through writing and polishing your business proposition until it actually says something, rather than sounding like soup. It’s a single-page app living at advisor.digisavvy.dev. Free to use, takes about 45 minutes, and it’ll get you somewhere a blank page never will.

GitHub PM tool

This one’s pure scratch-my-own-itch. GitHub is great for code, fine for issues, and genuinely bad at letting me look at client work the way my brain wants to. So I built a layer that sits on top of GitHub, pulls in the issues from all my configured repos, and lets me build my own views: by client, by request type, however I need to see it that day. It’s the “I refuse to switch project-management tools one more time” tool. Same data GitHub already has, just through my lens instead of theirs.

This tool replaced my longtime reliance on Zenhub – the only thing I miss from Zenhub is the sprint-level automation. Is there an issue in that repo to create it? You bet yo azz!

Agent Inbox

The newest one, and the one I’m most not-so-quietly proud of. It’s a self-hosted email setup (it can send and receive) that we point at client signup forms and payment forms to make sure the emails actually fire. You’d be shocked how often a form “works” but the confirmation email silently doesn’t, and nobody notices until a customer does. Now a little agent watches for it. Under the hood it’s a database, some file storage, and a job queue stitched together on Cloudflare’s free tier — the email-routing trick that makes it possible is half the reason I went down this whole rabbit hole in the first place.

Pike

Pike is the ambitious one. Pike connects to Missive, Gmail, Telegram, and Slack and is supposed to become my ultimate assistant — an on-demand agent that can spin up meeting requests on command. Supposed to. Right now, Pike earns his keep by organizing my inbox every morning and laying out my day in order of priority, which, honestly, is already worth the price of admission. He needs more training before he’s the J.A.R.V.I.S. I’ve got in my head. But we’re getting there, and I’m weirdly attached to them already. Pike named themselves, too… and Pike generated their own avatar… This is so dystopian.

Browser detection tool

So often, people hit a snag on their website, and it can be a chore to get information about the browser and OS from the person having the issue. So I just created a little website, a single-page app that detects that information and allows the person to send that information to me. I can trigger it on a visit, but I usually just have people fill out the info and send it to me. And again, it’s tied to Postmark and hosted on Cloudflare for free.

AI transfer agent

This is the one that made me sit up. We deployed an AI call agent (built on RetellAI) that answers the phone during the week, and I gotta be honest: I’m impressed. It’s already vetted a handful of callers and dropped the real ones straight onto my calendar, no human in the loop. It also, blessedly, shuts down the spam calls — you know the ones, the “we noticed your Google Maps listing needs attention” crowd. It’s good enough to deserve its own post, so I’ll do a proper write-up on it soon.

What’s next

A proposal tool. I’m tired of wrestling proposals together by hand, so this one will generate and edit them in my branding with my framework baked in, handle the basic management, and (the part I’m actually excited about) show me when someone opens one and how they poke around inside it, plus let them comment and review right there in the document. Proposals that talk back. More on that when it’s real instead of a list item in a notes file.

The bigger point

Five years ago, none of these were worth building. The time cost was brutal, the hosting was a hassle, and the smart move really was to rent a SaaS and move on. Today the building is fast and the hosting is basically free, and that completely changes the math on “should I just make the thing myself.” We’ve always believed in small, single-purpose tools over bloated ones. Now that belief is finally cheap to act on.

So here’s my nudge: you probably don’t need another monthly subscription. You might just need a scrappy little app that does the one thing that’s been bugging you. Go build it — or, if that sounds like a great idea you have zero interest in doing yourself, come talk to us and we’ll build it with you. Keep grinding.