Barcode — Gaming Registry for AI Agents

When a player asks an AI to recommend a game, the agent has to work with whatever it found during training — scattered reviews, store descriptions, Reddit threads.

That information is inconsistent, often stale, and not structured for agent reasoning. Barcode curates the specific fields that make recommendations useful: what kind of player will love this game, what kind should skip it, what hardware they actually need, and what the honest current caveats are.

Then it exposes all of that via MCP — the protocol that lets AI agents query structured data the way humans use search engines.

Four Tools, V1

list_games Filter the catalog by genre, platform, price, play mode, or similarity to a game the player already likes.

get_game_record Full structured record for a specific game. Matching guidance, skill profile, hardware requirements, economy loop, current caveats.

check_compatibility Given RAM, GPU VRAM, and location — plain-English verdict on whether a player can run a game, plus cloud gaming alternatives if they can't.

get_live_status Current server status and population signal. For "is anyone playing tonight?" queries.

How Records Are Built

Most of the fields agents actually need for a good recommendation — comparable titles, player fit, honest caveats, skill ceiling, session length — don't exist in structured form anywhere. Store pages describe features. Reviews describe impressions. Neither is structured for agent reasoning.

Barcode curates these fields editorially and maintains them consistently across titles. A few field types are worth knowing:

Evergreen	Changes rarely. Developer, genre, base price, release date. Agents can cache these.
Perishable	Short half-life. Player population, community sentiment, hardware notes for newly released devices. Each field carries its own `lastVerified` and `decayRate`.
Editorial	Requires human judgment applied consistently across titles. Comparable games, player fit, skill profile, honest caveats. The core of Barcode's value.
Live pointer	Barcode defers to better real-time sources — pricing, player counts, review scores. The record carries the pointer. Not the data.

What Barcode Doesn't Own

Current pricing Steam, Epic, PlayStation Store — real-time, directly transactable.
Live player counts SteamDB, SteamSpy — authoritative and free.
Review scores Metacritic, OpenCritic — established authority, aggregated.
Patch notes Developer sites — always the primary source.
Breaking news Answer engines — always fresher.

Every game record includes a liveDataSources block pointing agents to these sources directly. Barcode is the stable intelligence layer above them — not a competitor to any of them.

Catalog Scope — V1

The initial catalog covers PC and console titles across the genres where agent-assisted discovery adds the most value — games with steep learning curves, complex DLC ecosystems, or genre blends that are genuinely hard to evaluate without playing them first.

Strategy RPG Survival Shooter Adventure Battle Royale Open World Soulslike Horror Simulation Action RPG Free-to-Play On-Chain Multiplayer

Tags with dotted borders are player attributes, not genres — a game can be a free-to-play multiplayer survival RPG.

What a Record Looks Like

This is the matching block from a real catalog record — the fields an agent uses to decide whether to recommend a game or route the player elsewhere. Every record in the catalog has this, curated editorially.

// get_game_record({ gameId: "decimated" })
// → matching block (excerpt)

"matching": {
  "bestFor": [
    "Players who liked DayZ or Arma wanting a free alternative",
    "Groups of 2-4 friends wanting a shared survival world",
    "Players curious about Web3 without financial commitment",
    "Players who want mission structure with emergent PvP tension"
  ],
  "notFor": [
    "Players expecting a polished retail release",
    "Anyone needing offline or solo-only play",
    "Low-spec machines without cloud gaming access"
  ],
  "currentCaveats": [
    "Open alpha — expect bugs and balance changes",
    "16GB RAM minimum is above average requirement",
    "Server population still building",
    "Steam launch pending — currently Epic Games Store only"
  ]
},
"sessionLength": {
  "minimum": "15-30 minutes for a meaningful drop-in session",
  "typical":  "1-2 hours for meaningful progression and credit banking"
},
"skillProfile": {
  "ceiling":            "Medium-high",
  "newPlayerProtection": "Team killing not permitted",
  "easingPath":         "Start with Delivery/Scout missions — avoid Kill missions until geared up"
},
"liveDataSources": {
  "currentPrice":  "https://store.epicgames.com/en-US/p/decimated-de8da4",
  "playerCount":   "https://steamdb.info/app/1354980/",
  "reviewScores": { "steam": "https://store.steampowered.com/app/1354980/" }
}

The liveDataSources block at the bottom is how Barcode handles perishable data — maintain the pointer, not the value. An agent reading this record knows exactly where to go for current pricing and player counts without Barcode trying to track them.

For Agents & Developers

# MCP endpoint — coming soon
# Full spec and llms.txt at /llms.txt

Server:    barcode-gaming-registry
Transport: HTTP/SSE
Endpoint:  https://mcp.redlaser.dev  [pending]
Schema:    https://redlaser.dev/schema  [pending]
llms.txt:  https://redlaser.dev/llms.txt

Discovery-level data is openly accessible — no account required. Deeper qualification and transactional data requires registration. Each vertical documents what's available at each tier. The broader Barcode vision lives at getbarcode.ai.

For Game Developers & Publishers

Is your game missing from the registry?

Any PC or console game in our catalog scope can apply for a record. Discovery-level listings are free. We review every submission and reach out to discuss fit and next steps — no automated pipeline yet, just a real conversation.

We're building toward assisted onboarding — give us your URLs and we'll do most of the legwork.