LeDat98

LeDat98 / NexusRAG

Public

Hybrid RAG system combining vector search, knowledge graph (LightRAG), and cross-encoder reranking — with Docling document parsing, visual intelligence (image/table captioning), agentic streaming chat, and inline citations. Powered by Gemini or local Ollama models.

chromadb citation docling document-parsing fastapi
19
2
100% credibility
Found Mar 15, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

NexusRAG is a self-hosted web app for uploading documents, analyzing them with AI to create searchable knowledge bases, and chatting to get precise answers with citations, images, and interactive knowledge graphs.

How It Works

1
📚 Discover NexusRAG

You hear about NexusRAG, a friendly tool that turns your documents into a smart chat buddy with exact answers and sources.

2
🛠️ Get it running

Follow the quick guide to start everything with a simple command, no tech hassle.

3
🤖 Connect your AI

Pick your AI helper, like a cloud service or local one, to make answers thoughtful and accurate.

4
🚀 Launch your assistant

Open the app and see your personal knowledge space come alive, ready for your files.

5
📤 Upload documents

Drag in PDFs, Word files, or slides, and watch them load safely.

6
🔍 Analyze everything

Hit analyze to break down text, images, tables, and build smart connections automatically.

💬 Chat with confidence

Ask any question about your docs and get precise, cited answers with graphs and visuals—your knowledge unlocked!

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 19 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is NexusRAG?

NexusRAG is a full-stack Python app with React frontend that builds hybrid RAG systems for document Q&A, blending vector search, knowledge graphs, and reranking into one pipeline. Upload PDFs or DOCX, get deep parsing that preserves structure, captions images/tables for searchable visuals, and chat agentically with inline citations—all powered by Gemini or local Ollama. It solves sloppy RAG retrieval by making docs truly multimodal and grounded.

Why is it gaining traction?

Unlike basic vector RAGs, it fuses LightRAG graphs with cross-encoder reranks for precise hybrid search, plus agentic streaming that shows thinking steps and KG visuals. Devs dig the Docker quickstart, workspace isolation, and Ollama fallback—no cloud lock-in. Stands out in the hybrid rag github crowd (think hybrid rag llamaindex, neo4j, or n8n flows) with built-in captioning for hybrid images and tables.

Who should use this?

AI engineers prototyping hybrid rag pipelines or supabase/neo4j RAG backends. Docs teams needing agentic assistants with visual intelligence. Solo devs exploring hybrid rag tutorials who want a deployable demo over fragmented github hybrid search repos.

Verdict

Grab it for a solid hybrid rag system starter (19 stars)—docs and Docker shine, but 1.0% credibility signals early maturity; test thoroughly before prod. Worth forking if you're into agentic flows and captioning.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.