LiteParse is a lightweight, local CLI tool for extracting structured data and text from unstructured documents like PDFs, Office files, and images without cloud dependencies.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install liteparse-docs
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install liteparse-docs using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is LiteParse?
LiteParse is a high-performance utility designed to bridge the gap between unstructured documents and machine-readable data. By utilizing Openclaw Skills like LiteParse, developers can process a wide variety of formats—including PDF, DOCX, PPTX, and XLSX—completely offline. This tool is built for speed and privacy, ensuring sensitive data remains on your local machine while providing rich features like OCR and spatial bounding box detection.
Unlike cloud-heavy alternatives, LiteParse relies on proven local engines to handle complex document conversions. It is an essential component for building robust data pipelines where external API latency or privacy risks are a concern. Its ability to generate page screenshots and JSON metadata makes it a perfect companion for local AI agents and search indexers.
LiteParse Use Cases
- Rapid PDF text extraction for building local RAG (Retrieval-Augmented Generation) systems.
- Converting legacy Office documents into clean text or JSON for data analysis.
- Generating high-resolution page screenshots for visual LLM analysis.
- Batch processing large document repositories in air-gapped or secure environments.
- Extracting layout-aware data using precise bounding boxes for document reconstruction.
How LiteParse Works
- The user provides a single file or a directory path to the LiteParse CLI.
- The system identifies the file type and routes it to the appropriate local handler, such as LibreOffice for spreadsheets or ImageMagick for images.
- LiteParse extracts text content or performs OCR on scanned documents based on user-defined quality settings.
- The tool maps the extracted elements into a structured schema, optionally calculating bounding boxes for every text block.
- The processed data is output as a formatted string, a JSON file, or high-fidelity images for downstream use.
LiteParse Setup
LiteParse is easily installed via Homebrew and requires a few local dependencies to handle various file formats. Integrate it into your Openclaw Skills environment using the following steps:
# Install LiteParse via Homebrew
brew install llamaindex-liteparse
# Install LibreOffice for Word, Excel, and PowerPoint support
brew install --cask libreoffice
# Install ImageMagick for image processing support
brew install imagemagick
Verify your installation by running lit --version in your terminal.
LiteParse Data Schema & Taxonomy
LiteParse organizes its output to be highly machine-readable, especially when using the JSON format. This allows for precise data mapping in professional workflows.
| Data Component | Description |
|---|---|
| Text Content | Raw or cleaned string data extracted from the document body. |
| Bounding Boxes | Coordinate data (x, y, width, height) for visual elements and text blocks. |
| Metadata | Includes page counts, source format, DPI settings, and processing timestamps. |
| Screenshots | PNG or JPG renderings of specific pages for visual verification or AI vision tasks. |
LiteParse Advanced Features
- Recursive batch processing to handle entire directory trees with a single command.
- Custom configuration files (
liteparse.config.json) for maintaining consistent settings across projects. - Multilingual OCR support to handle documents in various languages using the tesseract engine.
- Adjustable DPI settings for balancing between processing speed and image/OCR quality.
- No-OCR toggle to significantly speed up processing for digital-native PDFs that don't require visual recognition.
SKILL.md
Loading
Related Openclaw Skills
SEO + GEO for SaaS
A complete SEO and Generative Engine Optimization system for SaaS companies to dominate search results and AI citations.
see-video
A powerful vision-based tool that extracts video frames into a structured grid for direct multimodal LLM analysis without proxy models.
Ma Huateng Thought OS
A strategic thinking framework that emulates Ma Huateng's perspective on product design, business boundaries, and ecosystem investment.
Aerospace Engineer Roadmap
A professional career guidance platform that generates personalized learning pathways and specialization roadmaps for aerospace engineers.
Dynamic Roster Copytrader
An automated Polymarket copytrading agent that dynamically ranks and follows top-performing wallets based on real-time PnL and win rates.
Huang Zheng Perspective Mindset OS
A cognitive operating system that analyzes business decisions through the lens of Huang Zheng's core principles like Benfen and long-termism.
Featured*
