Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/MuhammadSalmanAhmad/rag-pdf-highlighter/llms.txt

Use this file to discover all available pages before exploring further.

RAG PDF Highlighter is a Python package that locates text chunks inside PDF documents and returns an annotated copy with highlights applied. It ships as both a FastAPI microservice — ready to receive HTTP requests from any RAG pipeline — and a plain Python library you can call directly without starting a server. All inputs and outputs are compatible with LangChain Document objects, so it slots naturally into existing retrieval workflows.

​
Key features

3-tier text matching

Finds chunks using exact match, then sentence-level fallback, then collapsed-whitespace matching for character-spaced PDF artifacts.

Async PDF download

Downloads remote PDFs with httpx using non-blocking I/O, keeping the service responsive under concurrent load.

Stateless

Temporary files are cleaned up after every request. No state accumulates between calls.

Docker-ready

A single docker build and docker run command gets the service running in a container.

Library-friendly

The core highlight_chunks_in_pdf function raises plain Python exceptions. FastAPI is not required to use it.

LangChain Document compatible

Accepts langchain_core.documents.Document objects directly, with page_content and metadata.page fields.

​
Installation

Install the package from PyPI: 
pip install rag-pdf-highlighter
To install from source with development dependencies: 
git clone https://github.com/MuhammadSalmanAhmad/rag-pdf-highlighter.git
cd rag-pdf-highlighter
pip install -e ".[dev]"

​
Two ways to use it

As a microservice: Start the Uvicorn server and send POST /highlight requests with a PDF URL and a list of document chunks. The service downloads the PDF, applies highlights, and streams back the annotated file. This mode is suitable for multi-language stacks or teams that want a standalone service boundary. As a Python library: Import highlight_chunks_in_pdf directly and pass a local PDF path along with your Document list. No HTTP layer is involved. This mode is useful when your RAG pipeline is already written in Python and you want to avoid the overhead of a network hop.

​
Next steps

Quickstart

Run the service and send your first highlight request in three steps.

Guides

Learn how to deploy with Docker and integrate the library into your pipeline.

Build docs developers (and LLMs) love

Get started for freeTalk to us