Logseq Ask PDF

A Logseq plugin that allows you to ask questions about PDF highlights (text or images) using Large Language Models (LLMs).

A Logseq plugin that allows you to ask questions about PDF highlights (text or images) using Large Language Models (LLMs).

The demo video uses the paper Frans, K., Park, S., Abbeel, P., & Levine, S. (2024). Unsupervised Zero-Shot Reinforcement Learning via Functional Reward Encodings. as an example.



Background

The 2nd-brain app Logseq is a powerful tool for knowledge management.
It allows users to highlight the texts or images in the PDF and save them as blocks in Logseq.
But, Logseq lacks a feature that access the content of highlights directly. (because, highlights are just simply a block of uuid pointing the text or image in the PDF)

Logseq Ask PDF is a plugin that allows users to ask questions about PDF highlights directly within Logseq.
With this plugin, users don’t need to switch to the PDF file to find the highlights to ask questions.
This plugin also supports RAG (Retrieval Augmented Generation) to answer questions based on the content of the PDF.



Features

  • Ask questions about PDF highlights directly within Logseq
  • Supports asking both text and image highlights
  • Use the entire PDF as context
  • Seamless integration with Logseq’s UI
  • Compatible with OpenAI API and local models that are OpenAI API-compatible



Technical Details

  • Typescript
  • Langchain