LlamaIndex

LlamaIndex is a data framework connecting custom data sources to large language models, featuring event-driven Workflows, hybrid retrieval strategies, and production RAG pipelines.

When to Use

• Enterprise Document Ingestion & RAG: Parsing PDFs, Notion docs, Word files, and databases into optimized index structures.
• Complex Query Routing & Multi-Document Agents: Routing queries between vector search, SQL databases, and summary indexes.
• Hierarchical & Parent-Child Indexing: Preserving granular sentence chunks with links to full parent document contexts.
• Knowledge Graphs & Hybrid Search: Combining dense semantic embeddings with sparse BM25 keyword search.

Quick Start

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader

# Load documents from directory and build searchable in-memory index
documents = SimpleDirectoryReader("./data").load_data()
index = VectorStoreIndex.from_documents(documents)