Embeddable Game Standard

The Embeddable Game Standard (EGS) is an open protocol for building composable, provable on-chain games on Starknet. It defines how games expose their logic, how platforms mint and manage game tokens, and how frontends display real-time game state.

For a high-level overview of why EGS matters, see the EGS presentation deck.

The Problem

On-chain games are silos. Each game reinvents token minting, score tracking, settings management, and leaderboard logic. Platforms that want to embed multiple games must write custom integration code for each one. There's no standard way for a tournament system, a quest platform, or a social app to discover, launch, and validate games.

Three Pillars

EGS is built on three layers:

What You Can Build

Games that are automatically compatible with any EGS platform:

• Implement a single trait (IMinigameTokenData) to expose score and game-over status
• Optionally add settings (difficulty modes), objectives (achievements), and custom metadata
• Your game becomes mintable, trackable, and embeddable anywhere

Platforms that aggregate and orchestrate games:

• Mint game tokens on behalf of users with a single mint() call
• Receive automatic callbacks when scores update or games complete
• Run tournaments, quests, and reward systems across any EGS-compliant game

Frontends that display real-time game data:

• Query games, tokens, players, and activity through a unified API
• Subscribe to live score updates and game completions via WebSocket
• Decode packed token IDs client-side for instant display without RPC calls

Architecture at a Glance

Metagame ──→ MinigameToken (ERC721) ──→ Minigame
  │  ▲               │                      │
  │  │               └── Registry            ├── Settings (optional)
  │  │                                       └── Objectives (optional)
  │  └── IMetagameCallback (on_game_action, on_game_over, on_objective_complete)
  └── Context (optional)

Game Lifecycle: Setup → Mint → Play → Sync (update_game()) → Complete (game_over())

Learn more about the architecture →

AI-Assisted Development with Context7

All EGS repositories are indexed on Context7 and can be queried by AI coding assistants (Claude, Cursor, Copilot, etc.) for up-to-date documentation and code examples.

Claude Code

Context7 is available as an MCP server. Add it to your Claude Code config:

claude mcp add context7 -- npx -y @upstash/context7-mcp

Then ask questions referencing the library — for example: "Using /provable-games/game-components, how do I implement IMinigameTokenData?"

OpenAI Codex

Add Context7 to your codex.md or system prompt:

When working with EGS, fetch up-to-date docs via Context7 MCP using these library IDs:
- /provable-games/docs
- /provable-games/game-components
- /provable-games/denshokan
- /provable-games/denshokan-sdk

Other tools

Any tool that supports MCP servers (Cursor, Windsurf, etc.) can use Context7. See the Context7 docs for setup instructions.

Quick Links