PlayAtlas

A self-hosted recommendation engine that turns your Playnite library into an intelligent, searchable knowledge base.

Info

This documentation is a work in progress and may change at any time

Introduction

PlayAtlas is a self-hosted recommendation and classification engine designed to operate on top of a Playnite-managed gaming library.

It introduces a deterministic, versioned scoring system that evaluates games across multiple structural dimensions (e.g., Roguelike intensity, Survival mechanics, Narrative focus), transforming a static collection into an explainable, queryable knowledge space.

The system runs entirely within the user’s local network and does not depend on any third-party services.

Motivation

Modern storefronts are highly optimized for acquisition-driven discovery. They excel at recommending games you don’t yet own.

But a different question often goes unanswered:

From the 1,000+ games I already own, what should I play next?

Playnite already solves the hardest problem: aggregating and normalizing a user’s distributed game libraries into a single, unified source of truth.

PlayAtlas builds on top of that foundation, introducing a deterministic, multi-dimensional classification engine that transforms a static library into a structured space of explainable genre intensity and structural patterns.

Instead of recommending what to buy, PlayAtlas helps users rediscover what they already have.

It does so without relying on external services, opaque algorithms, or cloud infrastructure.