zensical/zensical
Zensical
Zensical is a documentation framework and markdown static site generator designed for authoring modular technical topics. It includes a static site build orchestrator that uses a dependency graph to parallelize asset compilation and perform differential updates.
The project features a federated search engine that aggregates results from multiple documentation projects using a combination of inverted indexes and vector-based semantic search. It also includes an LLM context generator to extract lean, semantic information from modular topics to optimize prompts for large language models.
The system manages technical documentation through versioned content trees, multilingual localization workflows, and project hierarchy management for complex layouts. Build processes are handled via task-based orchestration to prevent plugin interference, while content is managed as independent topics for reuse across different output channels.
Tools are provided to convert existing MkDocs configuration files and project structures to facilitate site migration.
Features
- Topic-Based Modularity - Structures documentation as independent modular topics to enable reuse across different versions and output channels.
- Technical Documentation - Organizes complex project hierarchies and versioned content trees into structured, searchable technical reference pages.
- AI Context Optimization - Generates lean, modular context files from documentation to optimize prompt accuracy and reduce token usage for LLMs.
- LLM Context Compilers - Extracts semantic information from modular topics to create optimized context files specifically for LLM consumption.
- Topic-Based Authoring - Structures content as independent topics to enable reuse across multiple output channels and different variants.
- Documentation Frameworks - Provides a framework for authoring modular technical topics that can be reused across different versions and output channels.
- Static Site Generators - Converts markdown files into responsive, multi-language websites with integrated search functionality.
- Federated Search Integrations - Aggregates search results from disparate documentation projects into a single unified federated index.
- Vector Search - Combines traditional inverted indexes with vector search to match user queries based on conceptual meaning.
- Build Orchestrators - Uses a task-based build engine and dependency graph to parallelize asset compilation and perform differential updates.
- Build Pipeline Orchestration - Coordinates discrete build tasks and parallel processing via a dependency graph to optimize site generation.
- Build and Task Automation - Breaks the build process into discrete tasks with explicit dependencies to prevent plugin interference.
- Differential Updates - Tracks content changes to trigger updates only for affected files, avoiding full project recompilation.
- Dependency-Aware Task Orchestration - Executes build steps by analyzing a dependency tree to parallelize operations and prevent plugin interference.
- Multilingual Content Mappings - Coordinates translation workflows by associating localized content with original source files through a structured directory system.
- Semantic Search Engines - Implements a search engine combining inverted indexes and vector search to match queries based on conceptual meaning.
- API Documentation Generators - Provides automated generation of cross-language technical reference pages and linked specifications across technology stacks.
- Differential Artifact Rebuilding - Rebuilds only the specific artifacts affected by content changes to ensure rapid updates for large documentation sets.
- Site Navigation Generators - Infers site structure and navigation automatically from content organization or programmatic logic.
- Development Preview Servers - Ships a high-performance web server with a pluggable middleware system for real-time site previews.
- Directory-Based Mapping - Coordinates translation workflows by associating localized content with original source files via a consistent directory structure.
- Parallel Build Processors - Distributes I/O and CPU-bound build tasks across available processor cores using a task graph.
- Project Hierarchies - Organizes complex layouts including monorepos and versioned content trees with dedicated build pipelines.
- Build Pipeline Extensions - Enables injecting or redefining build functionality using a module system with explicit contracts and language bindings.
- Versioned Documentation - Provides a mechanism to deploy and manage multiple versioned iterations of technical documentation based on branching strategies.
- Design System Foundations - Applies a brandable design foundation using a component system for alternative, theme-compatible layouts.
- Site Previews - Implements a high-performance web server with middleware to provide live previews of documentation changes.
- Frontend Asset Compilation - Minifies and compiles frontend assets during the build process to ensure immediate reflection of design changes.
- Localization Workflows - Manages localization using folder-based organization and automated draft generation for updated content.
