softwayer Softwayer is a software factory. Every product gets its own subdomain, its own database, its own billing — and two weeks to prove itself. Problem → prototype → ship → scale or kill. Nothing else ships. * Real means it costs money, burns hours, or breaks every week. If it doesn't, we don't build it.
Rule 01: if a problem doesn't cost more than $500 a month, we don't build for it. Move the sliders and find out where yours lands.
Annual cost of doing nothing
$0
Under $500 a month. Honestly? Keep doing it by hand — a tool would cost you more than the problem.
Four stations.
Fourteen days.
No decks.
Every product runs the same rail. Fail a station and it gets logged and killed. We keep the code, we drop the plans.
Cycle time avg · 11.4 daysIntake — the problem
It has to cost real money, happen again next month, and be solvable in under 800 lines. Example on file: “$2,000 a month to send 40,000 emails.” That qualifies.
STATION-01 · REV 03
Build — fourteen days
Days 1–2 research. Days 3–9 build. Days 10–12 we use it ourselves on real work. Days 13–14 ship the subdomain and turn on billing. Miss the date and it goes to the kill log.
STATION-02 · REV 03
Ship — its own subdomain
One tool, one subdomain, one database, one billing account. No shared monolith. If a product dies it dies alone and nothing else notices.
STATION-03 · REV 03
Scale or kill
Thirty days to reach 100 people who actually use it. If it doesn't, we archive the marketing, keep the source, and write down what we got wrong. Blunt, but it keeps the floor clear.
STATION-04 · REV 03
House rules
How we find,
and kill, ideas
43 ideas in the backlog. Six percent of them will ever ship.
Jan 04 — 09:14
Cold email platforms bill per contact. We sent 12,000 messages last month and paid $840 for the privilege. Wrote down: self-hosted sending, inbox rotation, flat price. Killed the per-contact pricing model in the spec on the first pass.
Jan 11 — 22:03
Scraper running on a proxy pool, 0.4 seconds a lead in testing. Ran a thousand records through it to check the shape of the data. Cost per lead came out roughly forty times cheaper than the API we were quoting against. Moved it to station 02.
Jan 19 — 16:41
Vault holds 1.2 million keys in the load test, 11ms at p99. That one stays. Rankpulse is another story — nobody on the test list opened it twice. If it doesn't move by the end of February it goes in the log and we move on.
Co-founder & CEO
Decides what goes on the line and what gets killed. Reads every problem submitted through this page and writes the reply himself.
Co-founder & CTO
Owns the build. Writes the code that ships each tool to its own subdomain — and the code that takes it back down when a unit fails a station.
Turns a recorded walkthrough into a written procedure your team will actually follow.
Payment received to a compliant invoice, emailed, in three clicks. Aimed at freelancers billing across borders.
Generate PDFs and ebooks at volume through an API instead of babysitting a headless browser.
Manifesto — v1.0
One problem, one subdomain, built until it is genuinely good at that one thing. No dashboards about dashboards. If it doesn't remove a bill you're already paying or hand back five hours a week, it doesn't leave the floor.
08 — Intake
Describe the thing you keep paying for, or keep doing by hand. If it clears rule 01 we'll write back with what we'd build and what it would take. If it doesn't, we'll tell you that too — usually faster.