finic-ai/rag-stack

🤖 Deploy a private ChatGPT alternative hosted within your VPC. 🔮 Connect it to your organization's knowledge base and use it as a corporate oracle. Supports open-source LLMs like Llama 2, Falcon, and GPT4All.

What it solves

RAGstack provides a self-hosted, private alternative to ChatGPT that allows organizations to chat with their own internal knowledge bases. By using Retrieval Augmented Generation (RAG), it enables users to query documents like PDFs or internal data sources without relying on external, public AI services.

How it works

The project acts as an orchestration layer that connects a user interface and server to a vector database and an open-source LLM. It uses Qdrant as the vector database to store and retrieve document embeddings. When a user asks a question, the system retrieves relevant information from the database and feeds it into an LLM (such as Llama 2, Falcon, or GPT4All) to generate a context-aware response.

Who it’s for

It is designed for developers and organizations that need to deploy private, secure AI infrastructure within their own Virtual Private Cloud (VPC) or local environment, ensuring that sensitive data remains under their control.

Highlights

  • Supports multiple open-source LLMs including Llama 2, Falcon-7b, and GPT4All.
  • Includes automated deployment scripts for major cloud providers: AWS, Google Cloud, and Azure.
  • Uses Qdrant for high-performance, self-hosted vector storage.
  • Provides a ready-to-use interface for PDF document ingestion and chat.

Related

  • Project
    frutik/awesome-search

    A curated, comprehensive knowledge base and index of concepts, tools, and research papers for building modern search applications and information retrieval systems.

  • Project
    CHIANGEL/Awesome-LLM-for-RecSys

    A curated collection of academic papers and resources exploring the integration of Large Language Models into recommender systems, categorized by their role in the recommendation pipeline.

  • Project
    shibing624/similarity

    A Java toolkit for computing similarity scores between words, phrases, sentences, and paragraphs, featuring sentiment analysis and Word2vec-based synonym discovery.

  • Project
    WeiboAI/VibeThinker

    A series of ultra-efficient reasoning models (1.5B and 3B parameters) that achieve frontier-level performance in math and coding through a Spectrum-to-Signal post-training methodology.