LlamaIndex

What is LlamaIndex?

LlamaIndex is an open source data framework for building applications that use large language models with private documents. It is made by LlamaIndex and combines an open source library with LlamaParse for document processing, LlamaExtract for structured extraction and LlamaCloud for managed indexes. The problem it solves is the gap between raw business files and useful AI answers: teams need to parse messy documents, retrieve the right passages and give an agent reliable context.

Key features

• Document parsing. LlamaParse handles PDFs, Office files, spreadsheets and images, and supports tables, charts, images, layout detection and structured JSON output.
• Wide format support. The parsing platform accepts 130 or more file formats, which covers most business documents without a separate conversion step.
• Structured extraction. LlamaExtract pulls fields from documents into a schema, with extraction targets per document, per page or per table row.
• Indexes for retrieval. LlamaCloud Index builds and manages retrieval indexes, with several index tiers, external data sources and webhooks for automation.
• Agent workflows. The open source library includes workflows and agent building blocks for combining retrieval, tools and model calls.
• Compliance options. The managed platform lists SOC 2, HIPAA and GDPR compliance, SSO, VPC deployment and custom business associate agreements on higher plans.

Who is LlamaIndex for?

Engineering and data teams building retrieval augmented generation, document search or document agents. It fits AI engineers who need a parsing layer for invoices, contracts, reports and technical manuals, platform teams that want a managed index instead of running a vector database and pipeline themselves, and organizations in finance, insurance, healthcare and manufacturing that work with large document sets.

Use cases

• Building a search assistant over technical manuals, policies or research reports.
• Extracting invoice fields, line items and totals into a structured database.
• Turning contracts into a question answering tool with citations back to the source page.
• Creating an agent that reads financial filings and compares figures across documents.

Pricing

LlamaParse has a Free plan at $0 per month with 10,000 credits and up to 100 users. Starter costs $50 per month and includes 40,000 credits, with pay-as-you-go usage up to 400,000 credits. Pro costs $500 per month and includes 400,000 credits, pay-as-you-go usage up to $5,000 per month, priority support and 100 users. Enterprise pricing is custom and adds volume discounts, higher rate limits, SSO and SaaS or hybrid deployment. Credits are priced at 1,000 credits for $1.25, and each parse, extract or index action uses a set number of credits. See the official LlamaIndex pricing page.

What's new

LlamaIndex published version 0.14.25 of its open source library on 21 September 2026. Read the 0.14.25 release.

Getting started

1. Create a free account at cloud.llamaindex.ai if you want to use the managed parsing and indexing platform.
2. Install the open source library with pip install llama-index.
3. Set your model provider key, then load documents and build an index or query engine.

pip install llama-index

Alternatives

• Langflow is a visual builder for LLM applications and agents.
• RAGFlow is an open source RAG engine built around document understanding.
• LangSmith is a platform for tracing, evaluating and monitoring LLM applications.

FAQ

Is LlamaIndex free?

The open source library is free to use. The managed LlamaParse platform has a free plan with 10,000 credits, then paid plans start at $50 per month.

What is LlamaIndex used for?

LlamaIndex is used to connect large language models to private documents. Teams use it for retrieval augmented generation, document search, structured extraction and document agents.

LlamaIndex vs LangChain?

LlamaIndex focuses on data ingestion, indexing and retrieval for LLM applications, while LangChain is a broader framework for chaining models, tools and agents. Many teams use both, with LlamaIndex handling document pipelines and another framework orchestrating the application.

Does LlamaIndex support PDFs?

Yes. LlamaParse supports PDFs along with Office documents, spreadsheets and images, and can extract tables, charts, images and structured JSON from them.

Yes. The open source Python library can run locally or on your own infrastructure. LlamaCloud adds managed parsing, extraction and indexes for teams that do not want to operate that pipeline themselves.