World analysis

The one-sentence reason this world exists. Without a premise the engine and a reader can't tell whether this is grimdark fantasy, a hopepunk solarpunk city, or a hard-sci-fi mining colony. Every other category gets evaluated against it.

Your premise lives on the world's logline + description. The engine checks both for length and for whether the premise carries any tension (a world WHERE x BUT y) so the rest of the report has something to compare against.

Whether the domains hang together. The engine flags missing foundations (a world without geography, culture, or history reads as a stage set), and worlds where 60%+ of entries sit in one domain (top-heavy worlds tend to read like one-trick settings).

Consistency sits in this section too: duplicate entry names, entries you flagged as 'demands testing' that haven't been settled yet, and contradictions across descriptions.

Climate, terrain, regions, water, light. The physical bones of your world. The engine checks not just whether you have entries here but whether the descriptions carry the kind of verbs and processes that make a place feel real (does the climate do anything to the people who live in it, or is it just a noun?).

Culture, religion, politics, social hierarchy, law and justice. Five categories grouped together because they almost always overlap in practice. The engine grades each individually but the radar shows them as a cluster so you can see at a glance whether your society reads multidimensional or single-axis.

Magic, technology, economy, trade, creatures and species. The systems that distinguish your world from the world outside the book. The engine pays special attention here to whether descriptions name costs, limits, and consequences (a magic system without a price reads as a wish-fulfillment lever rather than a constraint with stories in it).

Where this world has been (history), what's beyond it (cosmology), and what stories the people inside it tell themselves (mythology). The engine checks for years, wars, founding events, and named gods or legends. A world with no history reads as a snapshot; a world with no mythology reads atheistic in a way readers feel even when they can't name it.

How people speak and name things, and what an ordinary person does on a quiet Tuesday. The daily-life check is one of the most useful in the whole report: a world that doesn't describe food, work, home, family, and sleep is hard for prose to inhabit, even when every other category looks fine.

When you bind the run to a project, the engine pulls your project's open Promises and asks: does the world have entries that can pay these off? A promise about the cost of magic with no magic system entry that names a cost is an orphan promise, and the report flags it.

Conflict potential lives here too. The engine counts how many entries describe friction (rival, banned, scarce, threatened, opposed) and flags worlds where the friction ratio is so low the writer has nothing to push against.

Are entries more than placeholders? Do entries have at least a paragraph of description and a couple of concept tags, or are they nouns in a list?

Originality scans for stock-shape patterns like 'dark lord', 'chosen one', 'galactic empire', 'evil AI'. Stock tropes communicate fast but they don't differentiate. The engine flags worlds that lean on more than two without owning them deliberately.

Usability is the third leg here: does the world have the structured artifacts (maps, timelines, articles, calendars, languages) that the engine and AI need to be useful as context? A world with 50 World Index entries but no map is harder for the prose tools to lean on than a world with 12 entries, a map, and a timeline.