LlamaParse is a document parsing service that turns PDFs, Office files, spreadsheets and images into structured data that AI applications can use. It handles tables, charts, embedded images, multi-page tables and handwritten notes, and the page says it supports more than 130 file formats.
Beyond parsing, the platform includes schema-based extraction agents that pull defined fields from documents, classification and splitting based on natural-language rules, and a chunking and embedding pipeline for retrieval. Output can be Markdown, plain text, JSON, XLSX, HTML tables or an annotated PDF.
The page says it is built for businesses in finance, insurance, manufacturing and healthcare, and for teams building AI agents. Listed use cases include financial due diligence, invoice processing and searching technical documents.
You send documents through the web app or the API and choose a parsing mode. Basic parsing uses as little as 1 credit per page, and layout-aware agentic parsing uses more credits for higher accuracy. The team recommends testing new document types under different settings before setting up a persistent pipeline.
The free plan includes 10,000 credits a month, which the page says is roughly 1,000 pages. The Pro plan is listed at $500 a month, and pay-as-you-go is $1.50 per 1,000 credits. Enterprise pricing is custom and adds SSO and deployment in a private VPC.
The hosted version runs on a cloud tenant with data encrypted in transit and at rest. Cached data is kept for 48 hours unless caching is turned off. The page says LlamaParse is SOC 2 Type II, GDPR and HIPAA certified, and that it is available on the AWS and Microsoft Azure marketplaces.
LiteParse, a separate open-source tool, parses documents locally without an LLM.
