A cyber-occult interface for generating "hyperstitions" (fictional ideas that make themselves real) using a combination of Markov chains and grammar-based generation, fueled by user-generated entropy.

• Entropy Pool: Interactive particle system (Smoke/Warp physics) to seed randomness.
• Dual Engines:
  • Markov Chain: Probabilistic text generation based on a corpus of "Spirits" (Aspects).
  • Grammar Sigil: Template-based generation for structured prophecies.
• Aspects (Spirits): Toggleable source texts (e.g., Nick Land, Bible, CCru).
• Cyber-Occult UI: "OG Amber" monochrome aesthetic with grain/scanline effects.
• Batch Generation: Generate 50 outputs at once for quality testing.

graph TD
    subgraph "Ingestion (Python)"
        A[Raw PDFs/TXTs] -->|process_corpus.py| B(Structured JSON Corpus)
    end
    
    subgraph "HyperstitionEngine (React)"
        B -->|Load| C[Corpus Loader]
        C --> D[Markov Engine]
        C --> H[Grammar Engine]
        E[Entropy Pool <br/> p5.js] -->|Noise/Input| F[Entropy Store <br/> Zustand]
        F -->|Trigger| D
        F -->|Trigger| H
        D -->|Generated Text| G[Oracle Display]
        H -->|Generated Text| G
    end

Philosophy: We strictly avoid Large Language Models (LLMs) for the generation phase to preserve "divine randomness" and the chaotic nature of the cut-up technique.

For detailed setup and usage instructions, please see INSTRUCTIONS.md.

• Node.js (v18+)
• Python (3.9+) for corpus ingestion

1. Clone the repository.
2. Install frontend dependencies:

   npm install

3. Install ingestor dependencies:

   pip install -r ingestor/requirements.txt

1. Process your corpus (see Instructions).
2. Run the dev server:

   npm run dev

MIT