A robust PDF extraction API that transforms documents into structured, LLM-ready content including semantic AST, Markdown, and tables without page limits.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install frompdf-api
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 frompdf-api 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 frompdf?
frompdf is a high-performance extraction API specifically engineered for AI agents and modern LLM pipelines. While many tools struggle with complex formatting or strict page constraints, this addition to the ecosystem of Openclaw Skills enables developers to convert any PDF into a semantic Abstract Syntax Tree (AST). This means headings, paragraphs, tables, and lists are not just recovered as raw text but are properly typed and nested for programmatic use.
Designed to overcome the 20-page limit often found in native processing tools, frompdf offers a scalable solution for multi-hundred-page documents. It provides various output formats, including Markdown for human readability and specialized chunks for RAG ingestion, making it an essential utility for anyone building sophisticated document analysis workflows within Openclaw Skills.
frompdf Use Cases
- Processing massive technical manuals that exceed standard 20-page limits for AI analysis.
- Extracting complex tabular data from financial statements into structured JSON or Markdown.
- Performing semantic diffing between two versions of a document to identify meaningful content changes.
- Automating the preparation of document chunks for vector store ingestion and RAG pipelines.
- Assessing document complexity and accessibility through automated readability scoring.
How frompdf Works
- The AI agent or developer sends a PDF file to the frompdf API along with a preferred output format.
- The service processes the document, identifying structural elements like headers, tables, and metadata.
- The engine generates a semantic AST or the requested markup (Markdown, HTML, or Text).
- If specified, the content is split into LLM-ready chunks optimized for context window constraints.
- The structured data is returned via the API, allowing Openclaw Skills to perform downstream tasks like analysis or database storage.
frompdf Setup
To integrate frompdf with your Openclaw Skills, you first need to register for an API key. You can do this via a simple curl command which provides 10 free credits:
curl -s -X POST https://api.frompdf.dev/register \
-H "Content-Type: application/json" \
-d '{"email": "[email protected]", "password": "yourpassword"}'
Once you have your key, configure it in your environment to enable the skill:
export FROMPDF_API_KEY='frompdf_your_key_here'
frompdf Data Schema & Taxonomy
The skill organizes extracted data into a highly structured JSON format when using the default AST output. This structure is ideal for Openclaw Skills to parse programmatically.
| Element | Description |
|---|---|
title |
The document title extracted from metadata or content |
pages |
The total number of pages processed |
sections |
An array containing typed objects like heading, paragraph, table, or list |
type |
Defines the specific nature of the content element |
metadata |
Includes author, creation date, and other PDF properties |
Tables are specifically returned with headers and rows arrays, ensuring data integrity during extraction.
frompdf Advanced Features
- Semantic Diff: A unique endpoint to compare two PDFs and return structured changes rather than simple text diffs.
- LLM-Ready Chunks: Automatically pre-splits documents into optimized segments for vector database ingestion.
- Multi-Format Support: Seamlessly switch between JSON AST, Markdown, HTML, and plain text outputs.
- Encrypted Document Handling: Support for processing password-protected PDFs via API parameters.
- Readability Scoring: Built-in analysis to provide a 0-100 score on the document's reading ease.
SKILL.md
Loading
Related Openclaw Skills
Board WebMCP Skill
A bridge for AI agents to programmatically inspect, edit, and collaborate on visual boards and diagrams using the WebMCP protocol.
Agent Memory Kit
A structured memory framework that organizes AI agent knowledge into episodic, semantic, and procedural layers to prevent context loss.
Anatomy Quiz Master
An advanced AI-driven quiz engine designed to generate interactive anatomical assessments for medical students and educators.
Story
A local-first terminal utility designed for drafting, optimizing, and scheduling narrative content with total privacy.
Twitch
A local command-line utility for tracking Twitch channel data, scores, and leaderboards without a GUI.
AI Intelligent Budget Management
An AI-driven financial framework for multi-dimensional budget preparation, real-time execution tracking, and predictive variance analysis.
Featured*
