Licenses & tech stack
The Gaming Multiverse Reasoner, top to bottom. Two honesty notes first: the application and its data pipeline use zero third-party packages — everything is the Python standard library — and the browser UI is hand-written HTML/CSS/JavaScript with an inline SVG force layout. No frameworks, no CDN assets, no trackers.
This project's code
MIT License — Copyright (c) 2026 Mike Demopoulos. Source: github.com/Mike-Demo/gaming-multiverse-reasoner. Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files, to deal in the software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies, subject to including the copyright notice and this permission notice in all copies.
Open-source components
| Component | Role | License |
|---|---|---|
| FalkorDB | Graph database holding the 50,217-game multiverse graph; all path queries are Cypher run live against it | SSPL-1.0 |
| Redis | Storage engine FalkorDB is built on (RESP protocol, ACL reader user) | Tri-license: RSALv2 / SSPLv1 / AGPLv3 |
| Python 3.12 | Application runtime — standard library only (http.server), including the MCP server and the RESP client | PSF License 2.0 |
| IBM Granite 4 H Small | The narration model served by watsonx.ai (ibm/granite-4-h-small) | Apache-2.0 (model) |
| Docker Engine | Runs FalkorDB on the VPC virtual server and builds the app image | Apache-2.0 |
| Ubuntu 24.04 LTS | Operating system of the FalkorDB virtual server | Various free/open-source licenses |
| Debian (python:3.12-slim) | Base image of the application container | DFSG-free licenses |
| openCypher | The graph query language every path query is written in | Apache-2.0 (open specification) |
| Model Context Protocol | Agent interface exposed at POST /mcp | MIT (open specification) |
| JSON-RPC 2.0 | Wire format for the MCP server | Open specification |
Hosted services (not open source — stated plainly)
| Service | Role |
|---|---|
| IBM Code Engine | Hosts this web app (min-scale 1, always warm) |
| IBM VPC (bx2-4x16 VSI, us-south) | Hosts the FalkorDB container |
| IBM Cloud Object Storage | Frozen data snapshots (igdb-raw-2026-10-14-v2.tar.gz + manifest) |
| IBM Container Registry | Application container images |
| IBM watsonx.ai | Serves the Granite model; narration runs in a deployment space |
| IGDB API (Twitch) | Source of all game data — see attribution below |
| GitHub | Source hosting for the project repository |
Data & attribution
Game data comes from the IGDB API (Twitch Entertainment, Inc.) — titles, ratings, companies, genres, themes, engines, franchises, collections and similarity links for 50,217 games, pulled 14 Oct 2026 and frozen in IBM Cloud Object Storage. The QueerCade catalogue (a project by the same author) is merged in so all 753 of its games are present. Communities are computed by this project's own weighted label-propagation implementation (communities.py, MIT as part of this project) — no third-party graph libraries. Precomputed and inferred material is labelled as such in the UI; every path shown is computed live in FalkorDB at request time.
Built for the WeMakeDevs × FalkorDB “Graph Hacks” hackathon, 14–18 Oct 2026.