OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
RAG & Knowledge Systems

Nexusrag

Nexusrag is a self-hosted document knowledge-base application that parses files, combines vector search with LightRAG knowledge-graph retrieval, reranks passages with a cross-encoder, and provides streaming chat answers with inline citations.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Nexusrag — OpenViST project profile in RAG & Knowledge Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Nexusrag when you want to build and run a document-grounded chat interface that combines vector retrieval, knowledge-graph lookup and cited answers.

Who it is for

  • People building a searchable knowledge base from their own document collections.
  • Teams that need question-answering grounded in passages with page and heading references.
  • Developers configuring a RAG application with Gemini, Ollama or local sentence-transformers components.

Documented uses

  • Parse documents while retaining structural and page information.
  • Retrieve relevant passages through vector search and knowledge-graph entity lookup.
  • Rerank retrieved passages and generate streaming, cited answers.
  • Make extracted image and table descriptions discoverable through associated page text.
Before you choose

Requirements & limits

The repository is public and describes Nexusrag as open source. Its README displays an MIT license badge; the supplied repository evidence does not identify a license.

  • Gemini is the default provider and requires a Google AI API key; Ollama is an alternative for model use.
  • The documented stack includes Python 3.10+, React 19, FastAPI and Docker.
  • The README lists a minimum of 4 GB RAM and 5 GB disk, with a first Docker build downloading about 2.5 GB of models.
  • Docling is the default parser; Marker is configurable as an alternative, with differences in format coverage and GPU requirements.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Start the full stack with Docker

Set up the documented Docker deployment and open the local interface.

Before you start

  • Docker 20+
  • At least 4 GB RAM and 5 GB disk
  • A Google AI API key for Gemini, or configuration to use Ollama

Steps

  1. Clone the NexusRAG repository and change into its directory.

  2. Copy .env.example to .env.

  3. Edit .env to set GOOGLE_AI_API_KEY or switch the provider to Ollama.

  4. Run the command below and open http://localhost:5174.

    Documented command
    docker compose up -d

Expected outcome

The documented Docker quick start brings up the full stack for access through the local web interface.

Guide limitations

  • The README estimates the first build takes about five to ten minutes and downloads roughly 2.5 GB of machine-learning models.
  • Ollama requires configuring the provider and its server settings.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Vector search

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
NexusragParse documents while retaining structural and page information.Docker, WebThe repository is public and describes Nexusrag as open source. Its README displays an MIT license badge; the supplied repository evidence does not identify a license.
Chatbot with PDF for Semantic SearchPrepare document content as embeddings and save it to a supported vector database.WebThe repository is identified as open source and publicly visible on GitHub.
AI LLM PlaygroundAsk questions about PDF documents through an assistant interface.Web
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
LeDat98/NexusRAG · Official homepage title · retained in its source language
How we review sourcesSuggest a correction