OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
RAG & Knowledge Systems

Rhubarb

Rhubarb is an Apache-2.0 Python framework for building document and video understanding applications with multimodal models on Amazon Bedrock. It provides interfaces for question answering, summarization, data and entity extraction, classification and visual content analysis, and includes a FastMCP server for exposing these capabilities to MCP clients.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Rhubarb — OpenViST project profile in RAG & Knowledge Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Rhubarb when building a Python workflow that needs multimodal document or video analysis through Amazon Bedrock, particularly when outputs must feed an application or an MCP-compatible client.

Who it is for

  • Python developers building document-processing or video-analysis applications.
  • Teams integrating Amazon Bedrock model outputs into downstream applications.
  • Developers who want document and video analysis tools exposed through an MCP-compatible client.

Documented uses

  • Ask questions about local or S3-hosted documents.
  • Summarize documents, including selected pages, or stream summaries and chat responses.
  • Extract structured data, named entities and PII from documents.
  • Explain figures, charts and tables contained in documents.
  • Classify documents with multimodal embeddings.
Before you choose

Requirements & limits

The project is publicly available on GitHub under the Apache-2.0 license. The documented Python installation command is `pip install pyrhubarb`; use requires a boto3 session for Amazon Bedrock. Video analysis requires Amazon S3 storage.

  • Rhubarb is a Python framework for developers rather than a hosted document-analysis interface.
  • Amazon Bedrock is the model ecosystem; the README names Anthropic Claude multimodal and Amazon Nova models for document and video tasks, and Amazon Titan multimodal embeddings.
  • Document files can be local or in S3; video analysis requires S3.
  • The README lists PDF, TIFF, PNG, JPG and DOCX as supported document formats.
  • Documents longer than 20 pages can use 20-page sliding windows with an overlap setting from 1 to 10; document classification does not use this method and considers only the first 20 pages.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Ask a question about a local PDF

Use Rhubarb's Python document-analysis interface to submit a question about a local PDF.

Before you start

  • Rhubarb installed with pip as pyrhubarb.
  • Python and boto3.
  • A boto3 session configured for Amazon Bedrock access.
  • A local PDF file.

Steps

  1. Install the package with the command below.

    Documented command
    pip install pyrhubarb

  2. Create a boto3 session with boto3.Session().

  3. Import DocAnalysis and initialize it with the local PDF path and the session.

  4. Call run with a question in the message argument.

Expected outcome

The example returns a response to a question about the PDF, such as asking for an employee's name.

Guide limitations

  • The README example does not specify model selection or detail AWS account and model-access setup.
  • The documented supported document formats are PDF, TIFF, PNG, JPG and DOCX.
Documentation 1 ↗
Source-based practical guide

Ask a question about a video stored in S3

Submit an S3-hosted video to Rhubarb's video-analysis interface and request an analysis response.

Before you start

  • Rhubarb installed with pip as pyrhubarb.
  • Python and boto3.
  • A boto3 session configured for Amazon Bedrock access.
  • A video stored in Amazon S3.

Steps

  1. Install the package with the command below.

    Documented command
    pip install pyrhubarb

  2. Create a boto3 session.

  3. Import VideoAnalysis and initialize it with the S3 video path and the session.

  4. Call run with a question about the video and print the response.

Expected outcome

The example requests a response describing what is happening in the video.

Guide limitations

  • Video analysis requires S3 storage.
  • The README example does not specify model selection or detail AWS account and model-access setup.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
awslabs/rhubarb · Official homepage title · retained in its source language
How we review sourcesSuggest a correction