ColiVara

Active
GitHub Python NOASSERTION

Description

ColiVara is a suite of services for storing, searching, and retrieving documents based on visual embeddings. It uses vision models instead of chunking and text-processing, achieving state-of-the-art retrieval on both text and visual documents without OCR.

Key Features

  • Vision-based retrieval — uses ColPali vision language models for document embeddings instead of text chunking or OCR
  • 100+ format support — handles PDF, DOCX, PPTX, and 100+ other file formats with automatic image conversion
  • Metadata filtering — filter searches by collection and document metadata fields (author, year, tags, etc.)
  • No vector database required — managed PostgreSQL + pgVector backend handles all embedding storage
  • Python and TypeScript SDKs — full-featured SDKs with REST API and Swagger documentation
  • Late-interaction embeddings — outperforms pooled embeddings on both text-only and visual document benchmarks

Use Cases

💡 Enterprise document search — retrieving visually rich documents (charts, layouts, tables) that text-based search misses
💡 RAG systems — providing state-of-the-art document retrieval for retrieval-augmented generation pipelines
💡 Legal and compliance search — finding relevant clauses in contracts based on visual layout and structure
💡 Research paper retrieval — searching academic papers by figures, tables, and page layouts, not just text

Tags

rag vector-database embedding python data-processing

Categories

📚 RAG Tools

Quick Start

pip install colivara-py → from colivara_py import ColiVara → client = ColiVara(api_key='YOUR_KEY') → client.upsert_document(name='doc', document_url='https://example.com/file.pdf') → results = client.search('your query')
Visit GitHub Visit Website

Related Projects