Introduction

Rhubarb can perform multiple document processing and understanding tasks. Fundamentally, Rhubarb uses Multi-modal language models and multi-modal embedding models available via Amazon Bedrock to perform document extraction, summarization, Entity detection, Q&A and more. Rhubarb comes with built-in system prompts that makes it easy to use it for a number of different document understanding use-cases. You can customize Rhubarb by passing in your own system and user prompts. It supports deterministic JSON schema based output generation which makes it easy to integrate into downstream applications.

Features

  • Document Q&A

  • Streaming chat with documents (Q&A)

  • Document Summarization

    • Page level summary

    • Full summary

    • Summaries of specific pages

    • Streaming Summary

  • Extraction based on a JSON schema

    • Key-value extractions

    • Table extractions

  • Named entity recognition (NER)

    • With 50 built-in common entities

  • PII recognition with built-in entities

  • Figure and image understanding from documents

    • Explain charts, graphs, and figures

    • Perform table reasoning (as figures)

  • Document Classification with vector sampling using multi-modal embedding models

  • Auto generation of JSON schema from natural language prompts for document extraction

  • Logs token usage to help keep track of costs

Things to know

In it’s current form Rhubarb-

  1. Supports PDF, TIFF, PNG, JPG files. You can use upto 20 page PDF or TIFF documents

  2. Performs document to image conversion internally to work with the multi-modal models

  3. Works on local files or files stored in S3

  4. Supports specifying page numbers for multi-page documents

  5. Supports chat-history based chat for documents

  6. Supports streaming and non-streaming mode