hizi engine

hizi engine is a system for building and serving pre-computed game outcomes for fixed-odds games. Examples of fixed-odds games include slots, instant wins, and scratch cards. Every outcome it serves is certified and provably fair.

How It Works

+--------------+      +------------+      +--------------+
|  Your Game   |      |  hizi      |      |  JSONL       |
|  Engine      |----->|  Generator |----->|  (.jsonl.br) |
|  (offline)   |      |            |      |              |
+--------------+      +------------+      +------+-------+
                                                 |
                                                 v
+--------------+      +------------+      +--------------+
|  Your Game   |      |  hizi      |      |  hizi engine |
|  Frontend    |<---->|  RGS       |<---->|  (runtime)   |
|  (browser)   |      |  Platform  |      |              |
+--------------+      +------------+      +--------------+
        |
        +-- uses @hizi.io/engine-sdk

1. Generate: Your game engine simulates millions of outcomes. The hizi engine generator streams them into compact JSONL files (a line-delimited JSON format).
2. Deploy: You deploy the output files to the hizi RGS platform. The platform serves outcomes to players at runtime.
3. Play: Your frontend calls login, loadConfig, placeBet, and collect using the hizi engine SDK. The engine selects outcomes from the weighted entry pool. It uses certified RNG (a random number generator) for this selection.

What hizi engine Handles

• Outcome selection: The engine selects outcomes by weighted random selection, using certified RNG.
• Game state: The engine manages multi-step scenarios, free spins, and bonus rounds. It tracks the state across sequential placeBet calls.
• Balance management: The engine debits and credits the balance through the hizi RGS platform.
• Session management: The engine manages token-based authentication and the session lifecycle.
• Buy-feature pricing: The engine calculates pricing automatically from entry data.
• Certification: The engine verifies file hashes and keeps audit trails for RNG activity.
• Compact storage: The engine stores entries and scenarios as Brotli-compressed JSONL files. It serves outcomes from these files without memory accumulation, whatever the size of the dataset.

What You Build

• Game engine: Your game engine simulates outcomes offline (reels, wins, and bonuses).
• Frontend client: Your frontend client is the visual interface for the game (HTML5, React, Unity, etc.).

Next Steps

• Getting Started: Choose your path. Get set up.
• Core Concepts: Understand entries, weights, and scenarios.
• Provably Fair: The engine selects every outcome from a published, hash-verifiable entry pool. A player can verify a round later from the revealed seed.

For LLMs

This site publishes llms.txt index files. They follow the llmstxt.org standard. A large language model reads them easily.

• https://docs.hizi.io/llms.txt indexes every section of the site. Start here.
• https://docs.hizi.io/engine/llms.txt indexes this section only. Every other section publishes one too.
• https://docs.hizi.io/llms-full.txt holds the full text of every page, in one file.

Point your model, or your agent tooling, at the first of those to get a map of the docs.