ColiVara: COntextualized Late Interaction Vision Augmented Retrieval API

Elevate your document retrieval capabilities with ColiVara, a cutting-edge suite designed to store, search, and retrieve documents through innovative visual embeddings. Built as a web-first implementation of the ColPali paper utilizing the advanced ColQwen2 model, ColiVara transforms the user experience typically associated with traditional Retrieval Augmented Generation (RAG). Instead of relying solely on text processing, it taps into the rich visual content of documents, ensuring superior performance in retrieving information.

Quick Start Overview

To begin harnessing the full power of ColiVara, follow these steps:

1. Obtain a free API Key from the ColiVara Website.
2. Install the Python SDK for seamless API interaction:

   pip install colivara-py  
   

3. Easily index various document formats, including PDF, DOCX, PPTX, and more. ColiVara allows you to index using a file URL, base64 encoded string, or a file path, and even captures screenshots of webpages for indexing.

   from colivara_py import ColiVara  
   
   client = ColiVara(  
       api_key=os.environ.get("COLIVARA_API_KEY"),  
       base_url="https://api.colivara.com"  
   )  
   
   document = client.upsert_document(  
       name="sample_document",  
       url="https://example.com/sample.pdf",  
       metadata={"author": "John Doe"},  
       collection_name="user_1_collection",  
       wait=True  
   )  
   

4. Perform powerful searches with various filters to find documents efficiently:

   results = client.search("what is 1+1?")  
   results = client.search("what is 1+1?", collection_name="user_1_collection")  
   results = client.search(  
       "what is 1+1?",  
       query_filter={  
           "on": "document",  
           "key": "author",  
           "value": "John Doe",  
           "lookup": "key_lookup",  
       }  
   )  
   print(results)  
   

Key Features

• Advanced Retrieval: Utilizes state-of-the-art technologies to deliver superior retrieval quality and reduced latency.
• Comprehensive User Management: Enables multi-user configurations, allowing each user to manage their own collections and documents effortlessly.
• Extensive Format Compatibility: Supports over 100 file formats and automatically processes webpages, making it a flexible solution for various document types.
• Smart Filtering Options: Offers enhanced filtering capabilities by collection name, metadata, and custom fields.
• Convention Over Configuration: Designed for ease of use with optimized defaults, streamlining the development process.
• Flexible Components: Modular architecture allows integration with various storage solutions and APIs.

Documentation and API Access

Detailed documentation is available at docs.colivara.com. For those who prefer a visual interface, the ColiVara API Swagger documentation can be accessed here.

Why ColiVara?

Conventional RAG systems often fall short in processing visually rich documents due to their reliance on text extraction. ColiVara overcomes this limitation by employing vision models that create embeddings accommodating both textual and visual elements. As highlighted in the ColPali paper, ColiVara enhances information retrieval, making it suitable for a variety of applications where document design plays a crucial role.

_Join the community and explore the future of document retrieval with ColiVara!