Scry is a market-intelligence tool. It describes where a game sits against the existing market — it does not predict commercial outcomes and it does not give financial or business advice.
Every number in Scry traces to a public data source — Steam Store, Steam Reviews, SteamSpy, Steam News — or a transparently computed proxy on top of that data (tag-affinity, percentile bands, opportunity scores). Narratives are generated by a large language model over the same inputs a user can see; the LLM polishes voice but is not permitted to invent numbers.
None of this predicts whether your game will succeed. Market neighborhoods are descriptive: they tell you who's there and what the terrain looks like. The decision of whether to ship into that neighborhood remains yours.
Owner counts come from SteamSpy's public bands (nominal ±40% accuracy). Revenue estimates are derived from owner midpoints × list price, without discount or refund adjustment. Review-based positivity uses Steam's own cohort conventions. Every derived number has its own methodology page linked from the metric's display — follow those links for the formula, the data source, and the limitations.
The market-read paragraph, comparable analysis sentences, and other text-heavy surfaces are produced by a language model (Claude Sonnet at time of writing). The prompts are templated over structured data and the model is instructed not to invent numbers, but LLMs can still mis-summarize, miss nuance, or repeat stale training data. Treat narratives as a starting read, not a verdict — and check the underlying numbers on the same page.
Scry does not provide financial, investment, legal, or business advice. Nothing in the tool should be read as a recommendation to build, fund, publish, or abandon any project. Greenlight decisions and commercial bets belong to the people making them, informed by whatever combination of sources they find trustworthy — Scry is one of those sources, not the only one.