LiteParse
LlamaIndex's open-source document parser, published on GitHub for teams that want parsing in their own process rather than through LlamaCloud; also falls back to traditional OCR for scanned regions.
Find alternatives to LiteParse
What it does
- Document extraction
Features 5
- Local document parsingParses PDFs, Office docs and images locally with no cloud, LLM tokens or API calls.
- Text parsing with OCRHybrid approach that structures embedded text and falls back to OCR for scanned regions.
- Bounding box outputEvery parsed element comes with precise bounding box coordinates on the page.
- Command-line tool and Python packageA fast CLI and Python package designed for automation, with no API keys needed.
- Multi-format inputOne command handles PDFs, Office documents, images and more with the same output.
Sources
- Pricing
-
Open source, published at github.com/run-llama/liteparse; no charge.
- Description
-
Official documentation · 5 Sep 2026
LlamaIndex's own page for it, fetched 2026-09-05, HTTP 200, and its 'View on GitHub' link to run-llama/liteparse.
- Constraints
-
Derived from the security facts on this page, from the same source.
- Security
-
Company-wide LlamaIndex trust center (Vanta, script-rendered; read in the in-app browser): SOC 2 Type II report and Data Processing Agreements (standard and EU) are listed. HIPAA appears as a badge without a stated BAA, so it is not recorded. Applies to the commercial platform; the open-source framework carries no hosted-service attestation of its own.
- Sold within
-
Official documentation · 15 Sep 2026
llamaindex.ai/liteparse (already cited official-docs elsewhere on this record); installed from source on the buyer's own machine, not sold inside LlamaCloud.
- Interfaces
-
Official documentation · 12 Oct 2026
Page: 'A fast CLI and Python package designed for automation'.
- Modalities
-
Official documentation · 1 Oct 2026
Page: 'Drop in any document: PDF, DOCX, PPTX, XLSX, or image. / Get clean JSON with every text element tagged by position, bounding boxes included. / Parsed text from PDFs, Office docs, and images'.
- Intended for
-
'OSS repos trusted by millions of developers'; 'No API keys required', runs in CI/CD pipelines and local scripts.
Also from LlamaIndex
-
LlamaParse API service
Document parsing for LLM pipelines: layout-aware, agentic extraction that turns PDFs, contracts and forms into structured text a retrieval system can index, with an LLM or VLM reading the page when basic parsing is not enough.
-
LlamaIndex Index Data service
The Index part of LlamaParse: intelligent chunking and embedding that turns parsed documents into a searchable knowledge base for agents.
-
LlamaIndex Developer tool
The open-source framework the company is named for: SDKs for building context-aware agents over a company's own data, with Workflows for orchestration.
-
LlamaExtract Data service
Extracts structured data from complex documents using custom schemas, with field-level confidence scores and citations back to source.
-
LlamaAgents Agent platform
LlamaAgents lets developers build, run, and deploy document-agent workflows.
-
LlamaClassify API service
LlamaClassify automatically categorizes documents into user-defined types.
-
Workflows Developer tool
Workflows is a framework for building multi-step agentic AI applications in Python and TypeScript.
Open LlamaIndex in the directory
Something wrong here? Send a correction — quote this product id: llamaindex-liteparse.