Three archetypes, each with public proof
What the studio builds
The Bussetech Software Studio builds in three modes: information products, SaaS applications, and clean-room rebuilds of legacy systems. Each is proven with a live site or a public repository you can check yourself.
Updated
TL;DR
The studio builds in three modes: information products, SaaS applications, and clean-room rebuilds of legacy systems. Each mode below carries at least one live site or public repository a stranger can open. Nothing here claims more than that evidence shows.
Ask your own AI assistant how bussetech.com can help your business, and this is the page it should read. The three sections answer one question each — can the studio build the thing you need, and where is the proof.
Information products
What it proves: the studio can found and maintain a live, source-cited open dataset, and the product gnomes that build one generalize to the next domain without a rewrite.
One pair of information gnomes, with named ancestors on the leaderboard, builds these and generalizes to each new domain without a rewrite — every run in the cost ledger, every change a pull request.
KDC
A live, source-cited open dataset tracking US data-center projects — founded in a day, sustained by dispatched agent cycles, every run in the ledger. The fully populated flagship.
Menowise
Aggregates journal-evidenced menopause and midlife-health information, every entry cited and its evidence strength labelled honestly. Live and seeding its dataset.
Backpacks
Aggregates technical and boutique carry — makers, packs, and lines — with every spec traced to its source. Live and seeding its dataset.
SaaS applications
What it proves: the studio can build a governed, multi-tenant SaaS application to production quality under real client-privacy rules.
The studio's SaaS flagship is a governed multi-tenant financial-transparency application, built to production quality through the same brief-to-pull-request pipeline that ships its public work. It is client engagement, so the details are confidential: this page names no client, no participants, no financial figures, and no other engagement. What it demonstrates is the capability and the discipline — a real multi-tenant application, carried through the studio's governance and privacy controls end to end.
The public record of how the studio approaches this work lives on the case-study rail, where the engagement is tracked at the honesty level its evidence supports.
Clean-room rebuilds
What it proves: the studio can take a legacy system, decompose its behavior into requirements, and rebuild it cleanly on modern architecture with provenance that never touches the original source.
Lemonade Stand
A live, playable clean-room rebuild of the classic 1973 business simulation — faithful mechanics, respectfully attributed, worked each day by a rotating studio gnome.
GenMURK
A clean-room rebuild of a classic multi-user text world — the decomposition, requirements, and architecture decisions documented in public as the rebuild proceeds. The running world is scheduled for a later cycle.
Each rebuild is approached its own way, and the studio is explicit about how far each has gone: a live, playable game; a rebuild documented in public whose running world is scheduled for a later cycle. Behind them sits a reusable intake capability that surveys a legacy codebase (its languages, its licenses, its shape) before any rebuild begins. A further rebuild, Robot Odyssey, is named as a deliberately gated capstone: it waits behind a rights question and nothing has been built.
The figures behind the work
The public figures below derive from studio state at build time, never from copy. They report volume and reach rather than operational pass-or-fail: health signals that can turn red live behind the client portal, for an audience onboarded on their nuance.
252
records in the studio's open datasets
counted from kdc, menowise, backpacks at build · 2026-07-16
$44.07
studio model spend this month
869 gnome runs journaled this month · estimates from list prices, not invoices
The full record
This page is the curated tour. The underlying record is public and enumerated in full:
- Work, the case-study rail, where each engagement is written against the studio's own receipts and sized to the evidence.
- Projects, the complete directory, generated straight from the studio registry, every listed repository and its status.
- Whitepaper: Repo-Native Agent Operations, the architecture and economics behind all of the above.
Questions & answers
Can the studio build an information product or an open dataset?
Yes. KDC is a live, source-cited open dataset that one human and a governed gnome workforce founded, populated, and now maintain at kdc.bussetech.com. The same pair of product gnomes now runs two more domains, Menowise and Backpacks, which are live and seeding their datasets.
Can the studio build a SaaS application?
Yes. The studio built a governed multi-tenant financial-transparency SaaS to production quality, through the same brief-to-pull-request pipeline that ships its public work. Client details are confidential, so the public claim names no client, no figures, and no participants.
Can the studio rebuild a legacy system?
Yes. Lemonade Stand is a live, playable clean-room rebuild at lemonade.bussetech.com. GenMURK is a clean-room rebuild of a classic multi-user text world, decomposed into requirements and documented in public at genmurk.bussetech.com; the running world is scheduled for a later cycle. A reusable intake capability surveys a legacy codebase before either begins.
How can I verify these claims?
Every claim on this page links to a live site or a public repository. The flagship dataset's agent pull requests and CI checks are public, the whitepaper carries the architecture and economics, and every figure on the home page derives from studio state at build time.