About PS1 Discovery

PS1 Discovery is a search-and-browse tool over the entire PlayStation 1 library — 7,995 titles, deliberately weighted toward the obscure and the Japan-exclusive. The aim is to find a game by feel: you half-remember a mood, a mechanic, a fragment of a scene, and the tool surfaces candidates.

The data

Every game has an original one-paragraph description, written from that game's own evidence and adversarially fact-checked, plus controlled-vocabulary tags across ten dimensions (genre, mood, setting, mechanic, cultural lineage, discovery, quality, difficulty, length, recommendation). Underneath sits cross-linked metadata from 30+ sources in 11 languages — ratings, playtimes, prices, release regions, speedruns, achievements, and links out to each source.

The corpus is published openly as the PS1 Discovery Corpus (CC BY-SA 4.0), also mirrored on Hugging Face and archived on Zenodo with a DOI.

How search works

Keyword search matches your words against titles, studios, descriptions, and tags. Semantic search ranks by meaning: descriptions are embedded as vectors ahead of time, your query is embedded in your browser, and the closest matches come back — so "a game where you send letters" can find the right title even when none of those words appear in it. Both run entirely in your browser; nothing you type is sent to a server.

On rights

Descriptions here are original and licensed CC BY-SA 4.0. Third-party text (Wikipedia excerpts aside, which are CC BY-SA and shown with attribution) is linked, not reproduced. Cover art and screenshots are not hosted; those rights remain with their publishers.

This is Phase 2 of a three-phase project: an open dataset (done), this browse + semantic-search site, and eventually a small domain-expert reasoning layer over the same corpus.