LlamaIndex
Free tierDocument OCR & AI agents that turn complex documents into structured, LLM-ready outputs in seconds
Free tier available·All audiences·Powered by Multiple (VLM-powered, proprietary agents)·API available·Open source
Key strengths
Industry-leading document parsing across 50+ file types including handwriting, tables, and chartsVLM-powered agentic OCR with auto-correction loops for high accuracy on messy or complex documentsOpen-source LiteParse for fully local, cloud-free document processingEnterprise-grade security with HIPAA, GDPR, and SOC2 compliance plus VPC deploymentEnd-to-end document agent workflows: parse, extract, classify, split, index, and retrieve
Free tier + paid plans
San Francisco, USA
Founded 2022
Self-hostable
No ratings yet
- RAG pipeline construction: Use LlamaParse's index endpoint to chunk, embed, and store parsed documents for high-precision retrieval-augmented generation applications.
- Multi-step document agents: Build durable agentic workflows that parse, classify, extract, and act on document content using LlamaIndex's Workflows framework.
- Schema-based structured extraction: Define JSON schemas to extract specific fields (e.g., invoice totals, contract dates) from unstructured documents without any model fine-tuning.
- Local/offline document processing: Integrate LiteParse into CI/CD pipelines or air-gapped environments for fast, dependency-free parsing with bounding box output.
- Legacy IDP replacement: Replace template-based Intelligent Document Processing systems with LLM-native pipelines that generalize across document layouts.
- Context injection for LLM agents: Pre-process complex PDFs, charts, and scanned files into clean structured context so LLMs can reason over enterprise documents with human-level precision.