📝 Getting Started with QDrant Loader

Welcome to QDrant Loader! This section will help you understand, install, and start using QDrant Loader effectively, whether you're a content creator, researcher, developer, or system administrator.

🎯 Start path

  1. Understand the project: What Is QDrant Loader
  2. Complete setup and first ingestion: Quick Start Guide
  3. Use install troubleshooting only if needed: installation.md
  4. Review starter config patterns: Basic Configuration

🧠 Core Concepts

Understanding these key concepts will help you use QDrant Loader effectively:

🔄 Data Ingestion Pipeline

QDrant Loader processes content through a multi-stage pipeline:

  1. Collection - Gathers content from configured data sources
  2. Conversion - Transforms files (PDFs, Office docs, images) to text
  3. Chunking - Splits content into optimal segments for search
  4. Embedding - Creates vector representations using LLM providers (OpenAI, Azure OpenAI, Ollama, OpenAI-compatible)
  5. Storage - Saves vectors and metadata to QDrant database

🏗️ Multi-Project Architecture

  • Projects - Logical groupings of related data sources
  • Global Configuration - Shared settings (LLM, chunking, QDrant)
  • Unified Collection - All projects stored in same QDrant collection for cross-project search
  • Workspace Mode - Recommended approach for organized project management

🔌 MCP Integration

Model Context Protocol (MCP) connects QDrant Loader to AI tools:

  • MCP Server - Provides search tools to AI applications
  • Transport Modes - Stdio (default) and HTTP for different use cases
  • Search Types - Semantic, hierarchy-aware, and attachment-focused search
  • Real-time - Streaming responses for fast AI interactions

📊 Data Sources

Supported Sources with intelligent processing:

  • Git - Repositories, branches, commit history, file filtering
  • Confluence - Pages, spaces, attachments, hierarchy preservation
  • JIRA - Issues, projects, comments, attachment processing
  • Local Files - Directories, glob patterns, recursive scanning
  • Public Docs - External documentation sites with CSS extraction

🔍 Search Intelligence

Advanced Search Capabilities:

  • Semantic Search - Understands meaning beyond keywords
  • Hierarchy Search - Respects document relationships and structure
  • Attachment Search - Finds files and their parent documents
  • Cross-Document - Discovers relationships between different sources

🛤️ Recommended Learning Path

For Everyone (20 minutes)

  1. What is QDrant Loader? (3 min) - Project overview
  2. Quick Start (10 min) - Hands-on setup

For Users (Additional 10 minutes)

  1. Installation Guide (5 min) - Detailed installation
  2. Basic Configuration (5 min) - Configuration essentials

Next Steps

After completing the getting started section:

🆘 Need Help?