Aurum VOS is the operating system behind Aurum’s ventures: a shared platform every company inherits, a foundry that turns an idea into a launched product, governed AI executives that carry the daily work, and a command centre that decides what to build next and what to stop.
The thesis
Building a product is now cheap. Choosing the right one, distributing it, operating it well and stopping it on time are not. A venture operating system treats those four as the product, and micro-SaaS, small software businesses one person and a digital team can run, as its unit of work.
Identity, billing, email, analytics, AI routing, observability and the design system are inherited. A venture writes its data model, its business logic, its unique workflows and its brand. Nothing else.
Every hour invested in the platform is saved by every venture that follows. Company fourteen starts where company three finished. No product is the moat; the cost of making the next one is.
Each micro-SaaS launches with growth built in and a threshold, committed before launch, to keep going or stop. Winners get more. Everything else is retired on schedule and its learning feeds the next pick.
The system
Each part exists so that the next company is cheaper, faster and safer to launch than the last one. None of them is customer-facing.
Shared platform
Versioned packages every venture inherits: identity and tenancy, billing, email, notifications, consent-aware analytics, AI provider routing, content, accounting connectors and observability. The platform provides the mechanism; the venture provides the policy.
Creation engine
A durable workflow takes an idea through research, market scoring, thesis, specification, architecture, repository generation, implementation, QA and security review, up to a preview a person approves. State lives in the workflow, not in a chat.
Digital executives
CTO, CPO, CMO, CFO, COO, QA and Security, governed as services rather than personas. Default deny, scoped to one venture, budgeted, and recorded on a tamper-evident audit trail. They propose; people approve what reaches production.
Portfolio console
One screen per portfolio decision: what to build next, whether an idea passed its gate, which channel actually converts, and what to stop. Read-only by construction. It observes and proposes; it never writes to a venture.
Distribution layer
Growth installed by default: search and answer-engine readiness, programmatic content, localisation, outbound, experiments and one identity across the portfolio, so a launch is a growth engine from its first day.
Written principles
The rules every venture is generated from: simplicity, international from the first commit, minimum cost, product-led growth, tested by default, separability. Versioned; when it changes, every venture receives the pull request.
Lifecycle
Every stage has inputs, required outputs, quality gates, a budget and an escalation path. The Foundry can run for days because the workflow, not a conversation, holds the state.
A market signal, a micro-SaaS proven elsewhere, a regulatory change, a pain worth paying to remove.
Market, competitors, customer evidence and regulation, written into the venture’s own wiki.
Who it is for, why now, why us, and the archetype that decides its channels and gates.
Business model, pricing per country, product scope, data model and the workflows that make it different.
Repository, infrastructure, tests, documentation and growth surfaces, generated from the platform.
Agents build against contracts, in parallel, without touching the same branch.
Unit and system tests on every package, AI-powered QA and a security review before any preview.
A working product on its own domain, measured against the thresholds committed before it existed.
A person approves. Production is never an agent’s decision.
Experiments run, numbers flow into Command, and the venture earns more investment or a clean sunset.
Humans approve launches, production changes and anything that touches money.
Constitution
Short enough to be remembered, precise enough to be generated from.
Manifesto
“Companies should be as cheap to make as software.”
Twelve articles on how we build companies in the age of cheap software, and what changes for the people who build them.
Read the manifesto