SnapHound

SnapHound is an AI-powered image and video search tool that enables efficient retrieval using deep learning models.

Features

• AI-driven image and text-based search using SIGCLIP
• SQLite-based storage for embeddings

Installation

pip install -e .

or

pip install git+https://github.com/jebin2/SnapHoundPy.git

Usage

from snaphoundpy import SnapHound

# Initialize with paths to index
sh = SnapHound(paths=['/path/to/images'])

# Search with text
results = sh.search_with_text("A dog playing in the park")
print(results)

# Search with an image
results = sh.search_with_image("/path/to/query.jpg")
print(results)

Requirements

• Python 3.8+
• torch
• transformers
• numpy
• faiss-cpu
• pillow
• python-dotenv