Skip to content
Birchville Coat of Arms
Birchville Scholarly Synthesis
Canonical Editions of Classical Indological Reference Works

📖 Sanskritkurs Payer — Documentation & Architecture

Scholarly digitization, modernization, and AI-powered translation of Prof. Dr. Alois Payer's Sanskrit Course across 38+ target languages.

⚡ Quick Access & Applications

🌐 Interactive Course (PWA)
Complete textbook featuring 60+ lessons, exercises, and script modules in a modern offline-capable web application.
➔ Launch Course (PWA)
🏛️ System Architecture
Distributed 3-node topology with workstation (`nike`), dedicated LLM server (`nyx`), and staging runner (`nataraja`).
➔ System Architecture
🤖 AI Translation Pipeline
Weg-B translation strategy with Translation Memory, Devanāgarī protection filters, and 3-stage model hierarchy.
➔ Translation Pipeline
💻 Source Code & Pipeline
Open-source repository, Markdown parser (`markdown-it-extensible`), Service Worker, and Docker setups on GitHub.
➔ GitHub Repository

🌟 About the Project

The Sanskrit course authored by Prof. Dr. Alois Payer is one of the most comprehensive and academically rigorous German-language resources for learning classical Sanskrit. It is distinguished by its meticulous grammatical structure, extensive exercise sets, and rich philological commentary.

The Sanskritkurs Payer project transitions this academic resource from static 1990s HTML manuscripts into a modern, high-performance digital publishing platform: - Algorithmic Cleanup & Standardization: Transforming raw HTML into structured, future-proof Markdown with custom typographical extensions (markdown-it-extensible). - Modern Web Architecture: An ultra-fast VitePress Progressive Web App (PWA) capable of running completely offline. - Autonomous AI Translation Pipeline: Multilingual expansion across 38+ target languages while strictly preserving Devanāgarī characters, IAST transliterations, and complex grammar tables.

🏛️ Core Artifacts & Deliverables

🗺️ Documentation Overview

  1. System Architecture:
    Distributed 3-node topology (Workstation nike.local, LLM engine nyx.local, Staging/Runner nataraja.local), Mermaid system diagrams, and concurrency constraints.
  2. AI Translation Pipeline:
    The Weg B translation strategy, Translation Memory (TM) with MD5 hashing, Devanāgarī protection filters, and 3-stage model hierarchy (Local, Sonnet, Gemini).
  3. Language Selection & Criteria:
    Empirical findings from the 248.3-hour GPU benchmark, the 3-tier exclusion framework, and the 38+ language matrix.
  4. PWA & Infrastructure:
    Service Worker caching strategies, VitePress multi-lingual setup, Docker staging, and release pipelines.
  5. Developer Guidelines:
    Devanāgarī/IAST typographical rules, grammar-box syntax conventions (grammar-box), QA gates, and code governance.
  6. Wiki Index:
    Direct index of all reference articles in the official GitHub Wiki.

📜 License & Provenance

  • Original Work: Prof. Dr. Alois Payer (Tübingen).
  • Digital Edition, AI Pipeline & Web App: MIT License (Copyright (c) 2026 Marco Demarmels / Birchville Foundation).
  • Referenced Projects:
  • AlexandriaSandwich — Hybrid OCR & Book Digitization Pipeline
  • macdonell — A Vedic Grammar for Students (1916) 1:1 Digital Typeset
  • boethlingk — Pāṇini's Grammar (1887) Edition Pipeline