Archive ready

I built a self-hosted RAG system that actually works — here's how to run it in one command - DEV Community

https://dev.to/primoco/i-built-a-self-hosted-rag-system-that-actually-works-heres-how-to-run-it-in-one-command-38p2

April 1, 2026 at 08:35 PM JST•The archive page, viewer, and downloads use this saved version.

April 1, 2026 at 08:35 PM JST·dev.to

Bundle the HTML, screenshot, summaries, and metadata into one ZIP file. Pro saves automatically start preparing the external RFC 3161 timestamp, and only unfinished records need one more preparation step before download.

Saved page

I built a self-hosted RAG system that actually works — here's how to run it in one command - DEV Community

Open the dedicated viewer to inspect the saved page with archive metadata pinned above it.

Original URLhttps://dev.to/primoco/i-built-a-self-hosted-rag-system-that-actually-works-heres-how-to-run-it-in-one-command-38p2

StartedApril 1, 2026 at 08:35 PM JST

This is a self-contained HTML copy with CSS and images embedded, so it still renders even if the original page disappears.

The dedicated viewer keeps the original URL and saved timestamp visible while you review the archived HTML.

About this pageAI generated

This page describes RAG Enterprise, a self-hosted RAG system built by Francesco Marchetti to overcome limitations of existing tools. It addresses real business constraints: GDPR compliance, on-premise requirements, and budget limitations. The system is 100% local with no external APIs or data leakage. The architecture uses React frontend, FastAPI backend with LangChain, Qdrant vector database, and Ollama for LLM inference. It supports 10,000+ documents, 29 languages, handles PDFs and DOCX files, and includes multi-user role-based access control. Everything runs containerized on Docker, requiring Ubuntu 20.04+, NVIDIA GPU with 8-16GB VRAM, and 16GB minimum RAM.

Screenshot

The full page can be captured up to 15,000px in height so you can review the complete page layout when needed.