Enables searching documents by their visual appearance, going beyond traditional text-based methods.
PixelRAG provides a Python library for visual retrieval-augmented generation. It renders web pages into screenshots and builds a searchable index, allowing you to find information based on how it looks. You can install it with `pip install pixelrag` and use commands like `pixelshot` to process pages.
Enables searching documents by their visual appearance, going beyond traditional text-based methods.
Developers and researchers looking for new ways to index and retrieve information from web content, particularly those working with visual data or multimodal AI.
Not enough data yet. Star history will appear after a few days of tracking.
Official codebase for PIXELRAG: Web Screenshots Beat Text for Retrieval-Augmented Generation
Yichuan Wang*,
Zhifei Li*,
Zirui Wang,
Paul Teiletche,
Lesheng Jin
Matei Zaharia†,
Joseph E. Gonzalez†,
Sewon Min†
* Equal contribution † Equal advising
Work done at Berkeley SkyLab & BAIR & Berkeley NLP
Search any document by how it looks, not just the text it contains.
