Curiosity, with
a paper trail.

Observe. Preserve. Understand.

A private collector captures results already available in the in-game Auction House browser. We preserve the original SavedVariables file in a local SQLite archive, then derive structured insights from those observations.

What a price means here

V0.1 compares the minimum browse price for the same item key, collector, market context, client build, and manually labelled scope. Complete query results with no skipped rows are eligible. We compare the earliest and latest observations available, and report the actual interval.

These are asking prices. They don’t show completed sales, actual demand, or the whole economy. A matching manual label does not prove that the filters were identical. Each insight carries limited confidence and exposes these limitations.

Beta, live and PTR observations stay separate, as do rulesets, declared economic regions and capture pools, client regions/realms, factions, Auction House labels and builds. Context is declared per capture. An unknown economic region never means all regions; a comparison stays within its explicitly declared pool and raw client context. Auction House sharing between realms, factions or regions remains unverified. Legacy captures without this context remain archived and are excluded from new comparisons.

Samples stay visibly separate

The sample edition uses synthetic observations to demonstrate the pipeline and design. Sample and real data are never combined in an insight. If no insight file is supplied, the website shows an empty state.

From field note to history

Raw observations remain local and private. Only reviewed derived insight files belong on the website. More collectors, professions, rankings, and historical records will follow evidence about what the client reliably exposes.

Resource locations and gathering

A reference-location story describes an identified source and version, with its coverage made clear. Reference coordinates do not establish current spawns or gathering rates. Stories about activity over time need actual observations and a stated interval. No gathering Insights are available in this edition yet.

Class and progression observations

Class, race and level observations begin as private encounter diagnostics. Travel, visibility and repeat sightings shape the sample; they cannot establish population shares or which class levels fastest.

Item facts and credits

Item panels show metadata recorded with the observation, including the full Forever item key. They do not fetch current equipment stats. Diamonds are original placeholders while game artwork is unavailable. Sample metadata is synthetic.

Typography uses self-hosted Rubik. Copyright 2015 The Rubik Project Authors, distributed under the SIL Open Font License 1.1.