About

An on-demand atlas of the AI-native economy.

startups.studio is a directory of every business worth building in the AI era — generated on demand from the same authoritative standards governments and businesses use to classify work, and composed into autonomous startups you can read, fork, or run.

What it is

A directory of generated startups, grounded in standards.

Every URL on this site is a coordinate — a path like /CompanyTypes/Accounting_Firm/Problems/Reconcile_Bank_Statements that pins a position in a 16-dimension ontology. Visit one and a Thing generates at that coordinate: a typed entity with a description, structured facets, and links into the rest of the graph. Relationships flow into the prompts as free-text Prompt Context, not foreign keys — so any coordinate, even one no one has visited before, generates on demand.

Underneath, a value-flow knowledge graph wires how every commodity is produced, consumed, and delivered. Edges are first-class verbs (produces, realizes, composes, uses) and nodes are every entity type in the spine — seeded from NAICS (industries), O*NET (occupations & tasks), APQC PCF (processes), UNSPSC(products & services), and GS1 (value-flow lifecycle). The cascade enriches the gaps; it never fabricates the spine.

Who we are

One brand, four public surfaces.

startups.studio is operated as a coherent set of public surfaces — each a capitalized domain wordmark — that share one model, one cascade, and one quality discipline. The atlas at /explore is the discovery layer; the canonical wordmarks below are the four end-user faces where the work ships.

The discipline

Capability-confident. Honest by default.

Every generated page is capability-confident: present-tense claims about what the AI-native economy can do, today. We never fabricate a customer, a metric, a press mention, a fake testimonial, or social proof we haven’t earned. The atlas is research; we describe the research, not a marketing pitch.

Every LLM generation we run is append-logged to durable storage at the moment of generation — gates and quality checks are metadata, never gatekeepers. A heuristic verdict never decides whether real output gets to exist. Read the methodology for the full pipeline.

An atlas of what the AI-native economy can build.