Local Document Retriever with MixedBread Embeddings

A document retrieval system using ChromaDB and MixedBread embeddings for efficient semantic search capabilities. This project focuses on local deployment with the mixedbread-ai/mxbai-embed-large-v1 model.

Features

• Local embedding generation using mixedbread-ai/mxbai-embed-large-v1
• Document preprocessing with markitdown
• Efficient document storage and retrieval using ChromaDB
• Metadata filtering support
• Fully offline capable

Prerequisites

• Python 3.8+
• Local copy of mixedbread-ai/mxbai-embed-large-v1 model
• Sufficient storage for document embeddings

Installation

1. Clone this repository:

git clone <repository-url>
cd document-retriever