LlamaIndex
Document parsing and extraction platform (LlamaParse) plus the open-source LlamaIndex framework for document agents
LlamaIndex is best known for its open-source framework for building LLM applications and agents over private data, and its homepage now centers on LlamaParse, a platform that turns unstructured documents into AI-ready context. Developers use it to parse, extract, classify and index documents and to deploy document agents.
Key features:
- Parse more than 130 file types with agentic, layout-aware OCR.
- Route text, tables and charts to task-specific expert agents with auto-correction loops.
- Extract structured data to a schema, with citations and confidence scores.
- Read handwriting, dense tables and charts in 80+ languages.
- Build document agents from natural language and deploy them from GitHub.
- Index documents for RAG with external data sources.
The framework is open source in Python and TypeScript, and the OCRia-1 parsing model can run in LlamaParse or a customer's environment. LlamaParse is SOC 2, HIPAA and GDPR compliant; the free plan includes 10,000 credits a month (about 1,000 pages), Starter is $50/month and Pro $500/month.