LlamaIndex is an open-source Python and TypeScript framework for connecting large language models to structured and unstructured data sources using retrieval-augmented generation (RAG) and agentic patterns. It is widely used by developers to build production AI applications that query documents, databases, and APIs. LlamaCloud, the managed cloud product, adds enterprise-grade document parsing (LlamaParse) supporting 50+ file types, automated indexing, and scalable retrieval infrastructure. The framework is free; LlamaCloud is usage-based. Key features: - Open-source RAG and agent orchestration framework (Python and TypeScript) - LlamaParse: document parsing for PDFs, spreadsheets, images, tables, and handwritten notes - LlamaCloud: managed indexing, retrieval, and parsing pipeline with SLA-backed infrastructure - Integration with 100+ data sources, vector stores, and LLM providers - Agentic query planning with multi-step reasoning and tool use - Credit-based pay-as-you-go model for cloud parsing and indexing
Open-source framework is free. LlamaCloud: Free tier (limited credits); Starter $50/month (40,000 credits); Pro and Enterprise custom. Credits billed at $0.00125 per page (standard) to $0.05625 per page (complex documents).