What is LlamaIndex?
LlamaIndex (formerly GPT Index) is a data framework for building LLM applications. It provides tools for ingesting, structuring, and querying data from various sources so your LLM can answer questions about it.
Key Features
- Data Connectors — Ingest from 100+ sources: PDFs, Word docs, Notion, Slack, databases, APIs, websites.
- Indexes — Structure data into forms LLMs can query: vector stores, knowledge graphs, keyword tables.
- Query Engines — High-level interfaces for Q&A over your data.
- Chat Engines — Multi-turn conversational interfaces with memory.
- Agents — LLM-powered agents that use tools and query engines.
When LlamaIndex > LangChain
LlamaIndex is purpose-built for data indexing and retrieval. If your primary goal is RAG (making an LLM answer questions about a large document corpus), LlamaIndex's indexing abstractions are more mature and flexible than LangChain's.
Getting Started
pip install llama-index
LlamaIndex supports all major LLM providers (OpenAI, Anthropic, Mistral, local via Ollama) and all major vector stores (Pinecone, Chroma, Qdrant, Weaviate).