Extract text from documents and convert to Markdown
LlamaIndex reader that converts PDFs, Word docs, PowerPoint, HTML, CSV, XML, JSON, and ZIP files to Markdown using Microsoft's MarkItDown library.
Why it matters
Convert diverse file formats (PDFs, Word docs, presentations, spreadsheets, HTML, and archives) into clean Markdown text for indexing, analysis, or downstream processing in LlamaIndex workflows.
Outcomes
What it gets done
Parse PDF documents and extract text content as Markdown
Convert PowerPoint presentations and Word documents to structured text
Extract data from CSV, JSON, XML, and HTML files
Process ZIP archives and text files for content ingestion
Install
Add it to your toolbox
Run in your project directory:
curl -fsSL https://spark.entire.vc/get/li-reader-readers-markitdown | bash Overview
LlamaIndex MarkItDown Reader Integration
LlamaIndex MarkItDown Reader Integration is a document loader that converts files from eight different formats into Markdown text. It wraps Microsoft's MarkItDown tool to extract content from PDFs, Word documents, PowerPoint presentations, HTML files, CSV, XML, JSON, plain text, and ZIP archives, returning LlamaIndex document objects ready for indexing and retrieval. Use this reader when building LlamaIndex applications that need to process documents in multiple formats, such as RAG systems ingesting corporate file shares or knowledge bases aggregating mixed content types. It provides a unified interface for extracting text from common office and web document formats.
What it does
LlamaIndex MarkItDown Reader Integration extracts text from multiple file formats and converts them to Markdown. It uses MarkItDown to handle document ingestion from PDFs, Word documents, PowerPoint presentations, HTML pages, CSV files, XML files, JSON files, plain text files, and ZIP archives.
Inputs and outputs
You provide file paths to supported document formats: .txt files (including text files without extensions), .csv, .xml, .json, .html, .pptx, .docx, .pdf, and .zip files. The reader returns LlamaIndex document objects containing the extracted text content formatted as Markdown.
How to use it
Install the integration via pip:
pip install llama-index-readers-markitdown
Load documents in your Python scripts:
from llama_index.readers.markitdown import MarkItDownReader
reader = MarkItDownReader()
documents = reader.load_data("presentation.pptx")
Integrations
This reader uses MarkItDown to convert documents to Markdown format.
Who it's for
This integration serves developers building LlamaIndex-based applications who need to process diverse document collections. It is particularly useful for data engineers creating document ingestion pipelines, AI application developers building knowledge bases from corporate file shares, and researchers aggregating content from multiple file formats into unified text corpora.
Source README
LlamaIndex MarkItDown Reader Integration
MarkItDown is a powerful tool that converts various file formats to Markdown.
llama-index-readers-markitdown is an integration that uses MarkItDown to extract text from various file formats, supporting:
- .txt files and text-based files without extension
- .csv, .xml and .json files
- HTML files (.html)
- Presentations (.pptx)
- Word documents (.docx)
- PDF documents (.pdf)
- ZIP files (.zip)
You can install it via:
pip install llama-index-readers-markitdown
And you can use it in your scripts as follows:
from llama_index.readers.markitdown import MarkItDownReader
reader = MarkItDownReader()
documents = reader.load_data("presentation.pptx")
FAQ
Common questions
Discussion
Questions & comments · 0
Sign In Sign in to leave a comment.