Overview
Welcome to the RAG-lite TS documentation hub. This directory contains comprehensive guides for all aspects of using RAG-lite TS.
Quick Start
New to RAG-lite TS? Start with the CLI Reference for installation and basic usage, then explore the guides below.
Documentation Guides
📚 Core Guides
-
CLI Reference - Complete command-line interface documentation
- All commands and options
- Usage examples and workflows
- MCP server integration
- NEW: Multimodal mode support with
--modeand automatic reranking parameters
-
API Reference - Comprehensive programmatic API documentation
- Simple constructors and factory patterns
- Core classes and methods
- Type definitions and interfaces
- Usage patterns and examples
- NEW: Chameleon Multimodal Architecture with UniversalEmbedder interface and polymorphic runtime
-
Unified Content System - Memory ingestion and format-adaptive retrieval
- Memory-based content ingestion for MCP integration
- Format-adaptive retrieval (file paths vs base64)
- Dual storage strategy and content management
- Performance optimization and troubleshooting
- NEW: Complete guide for modern AI workflow integration
-
Configuration Guide - Advanced configuration options
- Configuration file setup
- Environment variables
- Multi-environment deployment
- NEW: Mode persistence and multimodal configuration options
🔧 Specialized Guides
-
Model Selection Guide - Embedding models and performance
- Model comparison and benchmarks
- Selection criteria and use cases
- Switching between models
- NEW: Multimodal models (CLIP) with text and image support
-
Path Storage Strategies - Document path management
- Relative vs absolute paths
- Portability and URL generation
- Multi-environment deployment
-
Document Preprocessing - Content processing and optimization
- File type handling (MDX, PDF, DOCX)
- Preprocessing modes and configuration
- Content extraction strategies
- NEW: Multimodal content processing with image-to-text conversion and metadata extraction
-
Troubleshooting Guide - Common issues and solutions
- Error message explanations
- Performance optimization
- Debug mode and diagnostics
- NEW: Multimodal-specific troubleshooting and model compatibility issues
-
Unified Content Troubleshooting - Specialized troubleshooting for unified content system
- Memory ingestion issues and solutions
- Content retrieval problems and recovery
- Storage management and cleanup procedures
- Performance optimization for large content operations
- NEW: Complete diagnostic and recovery procedures
🎨 Multimodal Capabilities (NEW)
-
Multimodal Configuration - Multimodal mode setup and configuration
- Setting up multimodal mode with CLIP models
- Processing mixed content (text + images)
- Reranking strategies for multimodal content
- Mode persistence and automatic detection
-
Multimodal Tutorial - Step-by-step multimodal guide
- Getting started with multimodal mode
- Cross-modal search examples
- Content type filtering
- Best practices and workflows
-
MCP Server Multimodal Guide - MCP integration for multimodal content
- Multimodal MCP tools and capabilities
- Image content retrieval through MCP
- Cross-modal search via MCP interface
- Configuration and usage examples
📊 Technical References
- Embedding Models Comparison - Detailed model benchmarks
- Performance metrics and analysis
- System requirements
- Technical implementation details
- NEW: Multimodal model performance and capabilities matrix
Documentation Structure
By User Type
New Users:
- CLI Reference - Quick start and installation
- Configuration Guide - Setup options
- API Reference - Programmatic usage
Developers:
- API Reference - Programmatic usage
- Unified Content System - Memory ingestion and MCP integration
- Model Guide - Model selection
- Preprocessing Guide - Content handling
System Administrators:
- Configuration Guide - Environment setup
- Path Strategies - Deployment patterns
- Troubleshooting Guide - Issue resolution
By Use Case
Basic Document Search:
MCP Integration & Memory Ingestion:
Technical Documentation:
Multi-Environment Deployment:
Performance Optimization:
Integration Development:
Quick Reference
Common Tasks
| Task | Primary Guide | Supporting Guides |
|---|---|---|
| Install and setup | CLI Reference | Configuration Guide |
| Ingest documents | CLI Reference | Preprocessing |
| Memory ingestion (MCP) | Unified Content System | API Reference |
| Content retrieval | Unified Content System | Unified Content Troubleshooting |
| Migrate to unified system | Migration Guide | Unified Content System |
| Search documents | CLI Reference | Model Guide |
| Choose embedding model | Model Guide | Configuration |
| Configure for production | Configuration Guide | Path Strategies |
| Handle different file types | Preprocessing Guide | Troubleshooting |
| Optimize performance | Model Guide | Troubleshooting |
| Deploy across environments | Path Strategies | Configuration |
| Integrate with applications | API Reference | Unified Content System |
| Resolve issues | Troubleshooting Guide | All guides |
Configuration Quick Links
- Models: Model Guide → Configuration
- Paths: Path Strategies → Configuration
- Preprocessing: Preprocessing → Configuration
- Environment Variables: Configuration
Troubleshooting Quick Links
- Installation Issues: Troubleshooting
- Model Problems: Troubleshooting → Model Guide
- Search Quality: Troubleshooting → Model Guide
- Performance: Troubleshooting → Model Guide
- File Processing: Troubleshooting → Preprocessing
Contributing to Documentation
Documentation Standards
- Clear structure: Use consistent headings and organization
- Practical examples: Include working code and command examples
- Cross-references: Link to related sections and guides
- User-focused: Write for the user's goals and context
Updating Documentation
When updating RAG-lite TS:
- Update relevant guides for new features
- Add examples to demonstrate new functionality
- Update cross-references if structure changes
- Test all examples to ensure they work
- Update this index if adding new guides
Documentation Feedback
Found an issue or have suggestions? Please:
- Check the Troubleshooting Guide first
- Search existing GitHub issues
- Create a new issue with:
- Which guide has the problem
- What you expected vs what you found
- Suggestions for improvement
External Resources
- Main Repository - Source code and issues
- NPM Package - Package information
- transformers.js - Underlying ML library
- Hugging Face Models - Available embedding models
This documentation is designed to be comprehensive yet accessible. Start with the guide that matches your immediate needs, then explore related guides as you become more familiar with RAG-lite TS.