LlamaParse is a state-of-the-art document parsing skill designed to transform complex PDFs, images, and spreadsheets into structured, LLM-ready data.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install llamaparse
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 llamaparse 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 LlamaParse?
LlamaParse is a heavy-duty document extraction engine that bridges the gap between messy files and structured AI input. By leveraging the LlamaCloud API, this skill enables developers to handle over 130 document formats, including scanned PDFs and complex presentations, converting them into clean markdown, text, or layout-aware data structures.
Built specifically for high-fidelity RAG pipelines and AI agent workflows, LlamaParse excels at preserving document hierarchy. Whether you are dealing with multi-column layouts, embedded tables, or visual charts, this Openclaw Skills integration ensures that your AI agents receive the most accurate representation of the source material possible.
LlamaParse Use Cases
- Converting complex PDF layouts into structured markdown for RAG applications.
- Extracting financial data and rows from embedded tables within spreadsheets or scanned reports.
- Performing high-accuracy OCR on scanned images and documents using intelligent parsing tiers.
- Automating the extraction of chart data into structured formats for data analysis.
- Batch processing large directories of documents for large-scale knowledge base construction.
How LlamaParse Works
- The document is uploaded to the LlamaParse cloud service via a secure API request.
- The user specifies a processing tier (e.g., agentic_plus, agentic, fast) to balance accuracy and speed.
- The engine analyzes the document's structure, identifying headings, paragraphs, tables, and images.
- Intelligent agents apply OCR and layout analysis to extract text while maintaining physical logic.
- Structured results are returned to the local environment in the requested formats, such as markdown or JSON-like items.
LlamaParse Setup
To integrate this skill into your workflow, ensure you have a valid API key from LlamaCloud.
# Install the required Python package
pip install llama-cloud>=1.0
# Set your environment variable
export LLAMA_CLOUD_API_KEY=your_api_key_here
LlamaParse Data Schema & Taxonomy
LlamaParse organizes output based on the expand parameter, allowing for granular control over the data returned:
| View | Description | Access Path |
|---|---|---|
| markdown | Structured text with headings/tables | result.markdown.pages[i].markdown |
| text | Flattened, clean text per page | result.text.pages[i].text |
| items | Layout-aware element tree with b-boxes | result.items.pages[i].items |
| metadata | Document-level metadata | result.metadata |
| images | Presigned URLs for extracted images | result.images_content_metadata.images |
LlamaParse Advanced Features
- Agentic Custom Prompts: Guide the parsing logic using natural language instructions to extract specific data points like receipt totals.
- Specialized Chart Parsing: Automatically convert graphical charts into structured data rows for analytical processing.
- Tier-Based Processing: Choose from four distinct performance tiers ranging from fast basic parsing to high-accuracy agentic analysis.
- Visual Consistency: Capture and download page screenshots for layout verification or visual AI processing.
- Concurrency Control: Utilize semaphore-based batch processing to handle large document volumes without hitting rate limits.
SKILL.md
Loading
Related Openclaw Skills
Search Bot
A high-performance real-time web search skill that delivers ranked results and AI-generated summaries for up-to-date information retrieval.
Git Workflows Pro
An expert-level git management skill designed to handle complex history surgery, repository architecture, and emergency recovery with a safety-first philosophy.
OpenClaw Context Upgrade
A specialized utility to diagnose and expand the conversation context limits of AI agents up to 1 million tokens.
kkclaw
A sophisticated desktop visualization layer and maintenance toolkit designed to give your AI agent a physical presence with emotional and voice feedback.
Open Ontologies
A professional-grade toolkit for AI-native ontology engineering, enabling automated generation, validation, and lifecycle management of RDF and OWL knowledge graphs.
Predict.fun MCP
A professional-grade MCP server providing structured access to Predict.fun prediction market data, yield mechanics, and trader behavioral analysis via The Graph.
Featured*
