An AI-powered tool designed to transform scientific literature PDFs into structured, machine-readable Markdown and CSV data.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install sci-data-extractor
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 sci-data-extractor 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 Sci-Data-Extractor?
Sci-Data-Extractor is a specialized tool within the Openclaw Skills ecosystem that automates the extraction of complex information from academic papers. By combining high-precision OCR technologies like Mathpix with advanced Large Language Models such as Claude 3.5 Sonnet and GPT-4o, it provides researchers with a reliable way to digitize experimental data, formulas, and tables.
This skill is built for high-stakes research environments where accuracy is paramount. It allows users to define custom extraction rules or use pre-configured templates to parse thousands of pages of scientific literature into structured formats, making it a cornerstone for anyone leveraging Openclaw Skills for technical data management.
Sci-Data-Extractor Use Cases
- Building enzyme kinetics databases by extracting Km and Kcat values from biochemistry papers.
- Compiling experimental results, including sample sizes and p-values, for meta-analysis.
- Generating structured literature review tables including DOI, methodology, and key findings.
- Converting complex mathematical formulas and PDF tables into clean Markdown formats.
- Automating the identification of specific protein structures or chemical compounds using custom prompts.
How Sci-Data-Extractor Works
- The user initiates a request by providing a scientific PDF and selecting a data template or custom prompt.
- The system utilizes Mathpix OCR or PyMuPDF to extract raw text, ensuring that mathematical notations and table layouts are preserved.
- Extracted text is processed by a configured LLM via the Openclaw Skills framework to identify specific data fields.
- The AI maps the content to the requested schema, ensuring data types and structures are consistent.
- The final output is generated as a Markdown table or CSV file for integration into research workflows.
Sci-Data-Extractor Setup
Prerequisites
- Python 3.8+
- API keys for Anthropic or OpenAI
Installation Steps
- Environment Setup (using uv for speed):
curl -LsSf https://astral.sh/uv/install.sh | sh
uv venv
source .venv/bin/activate
uv pip install -r requirements.txt
- Configuration:
Copy the template and add your API credentials to the
.envfile:
cp .env.example .env
# Add your EXTRACTOR_API_KEY and EXTRACTOR_MODEL
- Validation:
python extractor.py --help
Sci-Data-Extractor Data Schema & Taxonomy
The skill organizes extracted information into standardized schemas. Below are the primary data fields for the built-in templates available in Openclaw Skills:
| Template | Key Fields |
|---|---|
| Enzyme Kinetics | Enzyme, Organism, Substrate, Km, Kcat, Temp, pH, Mutant, Cosubstrate |
| Experiment | Experiment, Condition, Result, Unit, Standard Deviation, p-value |
| Literature Review | Author, Year, Journal, Title, DOI, Key Findings, Methodology |
Output can be toggled between --format markdown (default) and --format csv.
Sci-Data-Extractor Advanced Features
- Custom Prompting: Use natural language to define unique extraction targets like PDB IDs or specific cell lines.
- High-Precision OCR: Optional Mathpix integration for superior recognition of complex mathematical formulas.
- Flexible Model Support: Compatible with Claude 3.5 Sonnet, GPT-4o, and other API-compatible models.
- Segmented Processing: Advanced logic to handle long scientific documents that exceed standard token limits.
- Batch Export: Generate results directly into research-ready CSV files for immediate data analysis within Openclaw Skills workflows.
SKILL.md
Loading
Related Openclaw Skills
Running R Analysis in Existing Projects
Maintain and enhance R-based data science projects by extending scripts, updating visualizations, and troubleshooting statistical models.
RStudio Research Agent
A comprehensive agent skill for automating R-based scientific research, project scaffolding, and publication-quality data visualization.
PubMed-Search
An AI-powered assistant for comprehensive searching, metadata retrieval, and deep analysis of PubMed biomedical literature.
Google Scholar Search
An AI-powered academic search tool that enables researchers to discover papers, filter by author or date, and extract detailed scholarly metadata.
Sci-Hub-Search
An AI-powered academic assistant for searching, retrieving metadata, and downloading full-text research papers via Sci-Hub mirrors.
Scientific Graphical Abstract Generator
An AI-driven tool that transforms research descriptions and data into publication-quality, editable SVG visualizations.
Featured*
