Production Ready · v2.2.1

Semantic Document Search
Private. Local. Self-Hosted.

Imagine a magic bookshelf that reads and understands every book, document, and note you put on it. When you have a question, you don’t have to search for keywords yourself — you just ask the bookshelf in plain English (like “How do I fix the printer?” or “What was our revenue last year?”), and it instantly hands you the exact page that matters.

100% local. 100% private.

No cloud uploads. No accounts. No snooping.
Your personal librarian that keeps your secrets safe.

Not a chatbot. Not a cloud service.

PGVectorRAG is a desktop application that runs entirely on your computer. It helps you search and explore your own documents by meaning, even when you don't remember the exact words.

For example, searching for "dog" can also surface documents about "puppy" — because meaning matters, not just the word.

How it works

  • 1 You put your books and papers on the shelf
  • 2 The shelf reads them once and remembers what's inside
  • 3 Next time, it only re-reads pages that changed — so it's very fast
  • 4 You ask questions in plain language and get the right pages back
✓ Works completely offline after setup — no internet required
Get Started (Simple Install) 👉 View Technical Details (GitHub)

Powerful Features — Still Local

Built for real personal document libraries (and teams who value privacy)

You don't need to understand these details to use the app.

🔒

100% Local & Private

Your data stays on your machine. No cloud uploads, no privacy risks. AI that respects your secrets.

📄

Multi-Format Support

PDF, DOC, DOCX, XLSX, PPTX, CSV, TXT, HTML, Markdown, YAML, scanned documents (via OCR), and web URLs (CLI/API only). Automatic legacy .doc conversion.

🔍

Hybrid Search

Combine vector similarity with full-text search. Configurable weights for optimal results.

⚡

Incremental Indexing

Smart updates: automatically detects changed files and only re-indexes what's new. Saves time and resources.

🖥️

Desktop & CLI Tools

Desktop app for Windows, macOS, and Linux. Plus Docker CLI tools for servers and automation.

🐳

Docker Ready

Pre-built images on GitHub Container Registry. One-line installation for all platforms.

🗄️

PostgreSQL + pgvector

Production-ready database with HNSW indexing, connection pooling, and health monitoring.

🔄

Bulk Operations

Preview, export, delete, and restore multiple documents. Full backup/restore with JSON export.

🏷️

Document Types

Organize with custom types (e.g., policy, resume, report) or leave blank. Rich metadata storage and filtering.

🧪

Comprehensive Testing

268 tests covering unit, integration, and end-to-end scenarios. Separate test database.

🃏

Wildcard Search

Flexible filtering. Find everything matching a document type with simple wildcard queries.

📷

Smart OCR Modes

Three modes: Auto (smart fallback), Skip (fast, no OCR), or Only OCR (scanned docs only). Perfect control over processing.

🔐

Encrypted PDF Detection

Password-protected PDFs are detected and listed for review — your batch continues uninterrupted. Click the button to see the list and copy paths.

Common Ways People Use It

For personal and professional document libraries.

🔍

Private Document Search

Find exact information within your documents instantly using natural language, without relying on external services.

📚

Knowledge Management

Build enterprise knowledge bases with semantic search across policies, procedures, and documentation.

🔬

Research & Analysis

Index research papers, reports, and datasets for quick semantic search and discovery.

💼

Document Management

Organize and search contracts, invoices, resumes, and business documents with metadata filtering.

🏠

Always-On Home Search

Install on your NAS (Synology, QNAP, TrueNAS) or a spare PC. Search from any device on your network — always available, always private.

Get Started in Seconds

One-line installation for all platforms — installs and launches the desktop app automatically

This is for power users. Most people just use the desktop app — it's easier.

Desktop App (Recommended)

Works on macOS (including Catalina) and Linux.

Run this command to install and launch the desktop app:

bash 
curl -fsSL https://raw.githubusercontent.com/valginer0/PGVectorRAGIndexer/main/bootstrap_desktop_app.sh | bash

What this command does:

  • Installs required Python dependencies
  • Sets up the local PostgreSQL + pgvector database (via Docker)
  • Installs the desktop GUI
  • Launches the desktop app automatically

Features:

  • Full document management with drag-and-drop
  • Built-in backup/restore
  • Local semantic search interface

Requires a Docker-compatible container runtime for the database backend.

bash 
curl -fsSL https://raw.githubusercontent.com/valginer0/PGVectorRAGIndexer/main/docker-run.sh | bash

Headless Docker (API-only):

  • Downloads and orchestrates the pre-configured Docker containers
  • Includes PostgreSQL (pgvector) and the API server ready-to-run
  • Automatically initializes the database schema for you
  • No GUI — use via REST API or CLI

For server deployments or power users who prefer the command line.

Desktop App (Recommended)

Native Windows application with one-click launch.

Run this command in PowerShell to install and launch the desktop app:

powershell 
irm https://raw.githubusercontent.com/valginer0/PGVectorRAGIndexer/main/bootstrap_desktop_app.ps1 | iex

What this command does:

  • Installs required Python dependencies
  • Sets up the local PostgreSQL + pgvector database (via Docker)
  • Installs the desktop GUI
  • Creates a desktop shortcut for easy launch
  • Launches the desktop app automatically

Features:

  • Full document management with drag-and-drop
  • Built-in backup/restore
  • Local semantic search interface

Requires a Docker-compatible container runtime for the database backend.

📖

Available for all platforms (Windows, macOS, Linux)

Desktop app on all three; Docker required for the DB.

For manual setup instructions, troubleshooting, and advanced configuration, consult the full installation guide.

View Full Installation Guide

Your Personal Search Engine

Works as a desktop app. Advanced users can also run it in other ways.

📁

Native File Management

Access your local files directly without copying to shared folders

🔒

Private & Secure

Everything runs locally. Your data never leaves your machine.

🏷️

Smart Organization

Custom document types with rich metadata filtering

🗑️

Full Control

Bulk operations, backups, restores, and undo capabilities

Install the Desktop App Quick Start
Upload tab showing folder selection, OCR modes, and document indexing options Upload failures dialog showing detailed error list with filter options Encrypted PDFs dialog showing password-protected files detected during upload Search tab with semantic query input and ranked document results Documents tab displaying indexed files with metadata and source locations Recent tab showing recently accessed documents and search history Manage tab with database management and export backup options Settings tab for configuring embedding models and preferences

Frequently Asked Questions

▸ Does PGVectorRAGIndexer include installation or support services?

No. PGVectorRAGIndexer is designed as a self-service product. The focus is on clear documentation and reliable local operation rather than ongoing services.

▸ What platforms are supported?

macOS, Windows, and Linux/Unix.

▸ What file types are supported?

PDFs, text, Markdown, code, and other text-based formats. OCR is used automatically when required.

▸ Does my data leave my machine?

No. All indexing and search runs locally.

For Teams & Organizations

Deploy privately on an internal machine or server

  • ✓ Keep sensitive documents off the cloud
  • ✓ Optional API for integration
  • ✓ Commercial licensing available
Team Deployment Commercial License

Developer Mode / Optional API

For power users and developers building custom integrations

The core engine includes a REST API and a Developer Web UI (local) for developers who prefer browser-based access.

By default it binds to localhost (your machine only). You can optionally bind to your LAN for home server or team deployments.

Open Web UI (local)

Works only after you install and run PGVectorRAG locally.

API Docs (local)

http://localhost:8000/docs

POST /search 
{
  "query": "machine learning algorithms",
  "top_k": 5,
  "use_hybrid": true,
  "filters": {
    "type": "research"
  }
}
POST /upload-and-index 
curl -X POST "http://localhost:8000/upload-and-index" \
  -F "file=@document.pdf" \
  -F "document_type=policy"
GET /documents 
curl "http://localhost:8000/documents?limit=10"

API Features

  • Interactive API documentation at /docs
  • CORS support for web applications
  • Async operations for better performance
  • Comprehensive error handling
  • Metadata discovery and filtering
  • Bulk operations with preview
View API Documentation

❤️ Support the Project

PGVectorRAGIndexer is free for personal, educational, and research use.

If this tool saves you time, helps you stay organized, or simply feels valuable, you're welcome to support its continued development.

  • Donations are entirely optional
  • No features are locked
  • No tracking, no accounts, no obligations

Support the project on GitHub Sponsors

Donations do not grant commercial usage rights.

Commercial use requires a commercial license.