Skip to content

ROADMAP.md

Latest commit

 

History

History
24 lines (17 loc) · 1.75 KB

ROADMAP.md

File metadata and controls

24 lines (17 loc) · 1.75 KB
Raw

Project Roadmap

This document outlines the planned features and improvements for Rocketnotes.

Future Features

1. PDF & OCR Content Ingestion

Goal: Allow users to upload PDF documents and make their content searchable and editable within the application. This will enable the import of academic papers, scanned documents, and other reference materials.

Proposed Workflow:

  1. Frontend Upload: The user uploads a PDF file through the webapp.
  2. Backend Processing: The file is sent to a new, dedicated Python handler (handler-ocr) in the handler-ai project.
  3. Content Extraction & Conversion: The handler will use a specialized tool to convert the PDF into clean, structured Markdown.
    • Primary Tool (Open-Source): Marker is the recommended tool for this task. It excels at converting PDFs to high-quality Markdown, preserving formatting like headers, lists, and tables. It can be self-hosted within our handler-ai service.
    • Alternative (Cloud-Based): Amazon Textract is a powerful, fully-managed AWS service ideal for enterprise-grade, scalable OCR. It integrates seamlessly with AWS Lambda and can handle complex documents, forms, and tables with high accuracy.
  4. Note Creation: The resulting Markdown text is saved as a new note in the system.
  5. Indexing: The new note's content is immediately indexed by the existing vector embedding system (handler_vector_embeddings) to make it available for semantic search.

Key Benefits:

  • Unlocks a vast amount of existing knowledge stored in PDFs.
  • Makes scanned, image-based documents fully searchable.
  • Integrates seamlessly with the existing note and search architecture.