For 90% of local file tasks, you don't need a specific download; the built-in SimpleDirectoryReader automatically detects and uses the right internal loaders for you.
There are hundreds of community-contributed loaders on LlamaHub. How to Download and Use a Loader
LlamaIndex serves as the essential bridge between your proprietary data and Large Language Models (LLMs). While the core library is powerful, its true strength lies in its ability to ingest diverse data formats—from PDFs and Notion pages to Slack channels and SQL databases. This is made possible through the LlamaHub ecosystem and the specific process of downloading loaders. What is a LlamaIndex Data Loader? llama index download loader
A data loader (or Reader) is a specialized module designed to parse a specific file type or data source and convert it into a standardized "Document" format that LlamaIndex can understand. Why You Need Them
Historically, LlamaIndex used a specific download_loader function. However, the ecosystem has evolved toward a more modular, pip-installable approach for better stability and dependency management. Method 1: The Modern Way (Recommended) For 90% of local file tasks, you don't
from llama_index.core import download_loader # This downloads the loader code into your local environment GoogleDocsReader = download_loader("GoogleDocsReader") loader = GoogleDocsReader() documents = loader.load_data(document_ids=['your_id_here']) Use code with caution. Top 5 Most Popular Loaders to Download Loader Name Source Type Local Files Best for local folders with PDFs, TXT, and DOCX. BeautifulSoupWebReader Scrapes and cleans text from any URL. SlackReader Syncs conversations for internal knowledge bases. NotionPageReader Integrates structured notes and database entries. SimpleMongoReader Queries NoSQL databases for RAG pipelines. Best Practices for Data Ingestion
They turn messy raw data into clean text and metadata. While the core library is powerful, its true
Since different loaders have different requirements, use a virtual environment to avoid version conflicts between readers.