AnswerDotAI/byaldi
A Python library providing a simplified API for multi-modal late-interaction retrieval using ColPali and ColQwen2 models.

Byaldi wraps ColPali and related multi-modal models to enable easy retrieval over complex documents with embedded images. It converts PDFs to images and indexes them using late-interaction embedding techniques, allowing developers to perform semantic search across multi-modal content with just a few lines of code. The library is built on the colpali-engine and aims to support all ColVLM models, with plans for HNSW indexing and quantization optimizations.
Frequently asked
- What is AnswerDotAI/byaldi?
- A Python library providing a simplified API for multi-modal late-interaction retrieval using ColPali and ColQwen2 models.
- Is byaldi open source?
- Yes — AnswerDotAI/byaldi is open source, released under the Apache-2.0 license.
- What language is byaldi written in?
- AnswerDotAI/byaldi is primarily written in Python.
- How popular is byaldi?
- AnswerDotAI/byaldi has 851 stars on GitHub.
- Where can I find byaldi?
- AnswerDotAI/byaldi is on GitHub at https://github.com/AnswerDotAI/byaldi.