A sophisticated tool that converts static slide images into editable PowerPoint presentations using OCR and neural inpainting.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install image2pptx
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 image2pptx 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 image2pptx?
image2pptx is a high-performance utility designed to bridge the gap between static graphic assets and editable presentation formats. By leveraging advanced machine learning models, this skill reconstructs the underlying structure of a slide from a single image. It doesn't just overlay text; it removes the original text from the background and replaces it with native, selectable PowerPoint text boxes.
As a standout addition to Openclaw Skills, image2pptx utilizes a multi-stage pipeline involving PaddleOCR for text detection and LAMA for neural inpainting. This ensures that the generated .pptx has a clean, professional background while maintaining the original text content, font scaling, and color detection for a seamless editing experience.
image2pptx Use Cases
- Reconstructing editable slides from flattened image exports or screenshots.
- Converting infographics and posters into PowerPoint format for easy content updates.
- Localizing multilingual slides by extracting and replacing text while preserving the visual layout.
- Cleaning up presentation archives where the original source files have been lost.
- Automating the conversion of slide-based training materials into editable corporate templates.
How image2pptx Works
- OCR Detection: The skill uses PaddleOCR PP-OCRv5 to identify text regions, bounding boxes, and specific content.
- Textmask Generation: Classical computer vision techniques find specific ink pixels via adaptive thresholding.
- Mask-Clip Processing: The textmask is combined with OCR bounding boxes to ensure non-text illustrations are preserved.
- Neural Inpainting: The LAMA model reconstructs the background pixels behind the text to create a clean, unobstructed slide.
- PPTX Assembly: The skill finally assembles the .pptx file, placing editable text boxes with auto-scaled fonts and detected colors over the inpainted background.
image2pptx Setup
To integrate this capability into your workflow and expand your Openclaw Skills library, install the package via pip:
git clone https://github.com/JadeLiu-tech/px-image2pptx.git
cd px-image2pptx
pip install -e ".[all]"
Models (approximately 370 MB) will be downloaded automatically upon the first execution.
image2pptx Data Schema & Taxonomy
The skill manages various data types during the conversion process to ensure high-fidelity outputs:
| Data Type | Description |
|---|---|
| Input Image | Supports standard formats like PNG and JPG (WebP should be converted to PNG). |
| Intermediate OCR | JSON metadata containing text coordinates and content (can be pre-supplied). |
| Model Cache | Local storage for OCR and LAMA weights in user home directories. |
| Output PPTX | Standard OpenXML PowerPoint file with editable shapes and text boxes. |
| Work Directory | Optional directory to save debug files and intermediate mask layers. |
image2pptx Advanced Features
- Neural Inpainting: Utilizes big-lama to intelligently fill in backgrounds where text used to be.
- Multilingual Support: Optimized for both English and Chinese (ch) language detection.
- Custom Sensitivity: Fine-tune textmask sensitivity and dilation to handle varying font thicknesses.
- Developer API: Direct Python integration via the
image_to_pptxfunction for automated pipelines. - Selective Processing: Ability to skip inpainting for solid backgrounds to increase processing speed.
SKILL.md
Loading
Related Openclaw Skills
osv-ui CVE Audit Skill
A comprehensive security auditing skill that scans for CVE vulnerabilities and provides a visual dashboard for managed dependency remediation.
Sector Charts
A visualization tool designed to map competitive positioning using multi-dimensional bubble charts for market analysis.
Firecrawl
Firecrawl is a high-performance web scraping and crawling engine designed to transform entire websites into clean, structured markdown for AI applications.
Seven: The Cultural & Scientific Intelligence of 7
A comprehensive bilingual knowledge skill providing deep cultural, scientific, and symbolic insights into the number 7, alongside the OpenClaw project matrix.
Gardening & Plant Growth Video Generation
A professional AI video generation tool specialized in creating cinematic gardening timelapses and plant growth cycles.
Quant Trading Assistant
An advanced A-share quantitative trading assistant providing real-time data, technical indicators, and automated decision support.
Featured*
