Steady Finch, ConsentBloom, and writing in public
I'm starting a cadence on this site: a post about every three days, and a weekly note on what moved and what I learned. Steady Finch is what I'm building. This is the map.
Steady Finch
I live and build from Zlín. Loy software s.r.o. — the Steady Finch company — is registered in Prague. I'm the only founder.
The plan is a portfolio of small niche SaaS products — not one giant platform brand in the customer's face. Each product should stay narrow. Underneath, I want a shared spine: auth, billing, email and SMS transport, evidence. The boring parts that every product needs, built once.
The first public company site exists. The first product is ConsentBloom.
Why ConsentBloom
The people I have in mind are solo US beauty pros — lash, brow, wax — whose booking tools don't include consent forms. The pain isn't "I need a form builder." It's the chase: send the form, remind the client, then find the signed copy later when you actually need it.
I looked at a few early ideas. Beauty consent ranked among the STRONG ones. I chose it as the first thing to ship.
Version one is a fixed template pack and an automated chase — invite plus reminders. Email-first. Appointments entered by hand. It should sit beside Square, Booksy, and the rest, not replace them. I'm not building a booking app, and I'm not wiring into theirs.
Pricing direction, from the docs I've already approved, is about $12 a month. That's enough to say for now.
What the monorepo actually looks like
The SteadyFinch repo is private and architecture-first. There's apps/ for the company website and the console, products/consent-bloom/, services/core/, and shared packages. The stack is Bun, Turborepo, Next, Neon, and Vercel.
Four deployables are in my head: steadyfinch.com, a console, an API, and consentbloom.com.
The honest recent picture: I've built more platform, docs, and delivery system than customer-facing ConsentBloom UI.
What exists today: owner magic-link sign-in, company trust and privacy pages, a lot of core email plumbing, and billing schemas. What is not shipped: the client signing page (/s), Polar checkout, SMS in production, and a real ConsentBloom marketing landing. That last one is still a placeholder title in the code.
I'm building the factory and the first product at the same time. I'm also trying to say so when the landing page isn't live yet.
What I'll write next
I'm aiming for a post about every three days, plus a weekly progress and lessons note. A few seed titles, if you want to skim what's coming:
- The booking apps already won — so we refused to be one
- We don't sell a form builder
- Email-first, because SMS is a compliance product
- Architecture before the landing page
- Polar now, checkout later
- Chasing is the product
- Building a factory for niche SaaS
- Weekly: what moved, what stuck, what I learned
There's an RSS feed if you want these in a reader. More soon. If something here is useful or wrong, write me at alois@belaska.me, or use the homepage links.