A comprehensive parsing engine that converts images, documents, and audio into structured, LLM-ready text using OCR and automated transcription.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install multimodal-parser
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 multimodal-parser 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 Multimodal Content Parser?
The Multimodal Content Parser is a powerful integration designed to simplify the data ingestion layer for AI agents. It provides a unified interface to handle diverse file formats including images, PDFs, Word documents, and audio files, eliminating the need to manage separate APIs for different media types. By leveraging this Openclaw Skills contribution, developers can easily transform unstructured data into clean text, Markdown, or structured JSON.
Built on top of industry-standard open-source engines like Tesseract and Whisper, this skill handles the heavy lifting of optical character recognition (OCR) and speech-to-text. It is specifically optimized for LLM contexts, ensuring that the output is formatted for high-quality prompt injection or knowledge base construction.
Multimodal Content Parser Use Cases
- Building content analysis layers for multimodal AI agents.
- Automating document pre-processing for RAG (Retrieval-Augmented Generation) and knowledge bases.
- Extracting text from images and scanned PDFs for archival or search.
- Transcribing meeting recordings and audio notes into structured summaries.
- Batch processing heterogeneous file sets into standardized JSON formats.
How Multimodal Content Parser Works
- The user provides a file path and optional parameters such as file type or output format.
- The system automatically detects the file type if not specified (supporting Image, PDF, DOCX, or Audio).
- Depending on the format, the skill triggers the relevant engine: Tesseract for OCR, Poppler for PDFs, Pandoc for DOCX, or Whisper for audio.
- The raw data is processed and converted according to the requested output format (text, markdown, or structured JSON).
- The skill returns a standardized response object ready for consumption by an LLM or database.
Multimodal Content Parser Setup
To use this multimodal parser within the Openclaw Skills ecosystem, ensure the required system dependencies are installed for your environment.
# For macOS users
brew install tesseract tesseract-lang poppler pandoc
pip install openai-whisper ffmpeg
# For Ubuntu/Debian users
apt install tesseract-ocr tesseract-ocr-chi-sim poppler-utils pandoc ffmpeg
pip install openai-whisper
Multimodal Content Parser Data Schema & Taxonomy
The skill accepts several configuration parameters to fine-tune the parsing logic:
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
| file_path | string | Yes | - | The local path to the file being parsed. |
| file_type | string | No | auto | image, pdf, docx, audio, or auto detection. |
| output_format | string | No | text | Format of result: text, markdown, or structured. |
| options.ocr_lang | string | No | chi_sim+eng | Target languages for OCR recognition. |
| options.audio_model | string | No | base | Whisper model size (base, small, medium, large). |
| options.pdf_page_range | array | No | - | Page range for PDF parsing, e.g., [1, 10]. |
Multimodal Content Parser Advanced Features
- Automatic file type detection for zero-config integration.
- Granular PDF control allowing for specific page range extraction to save tokens.
- Adjustable Whisper model sizes to balance transcription speed and accuracy.
- Modular architecture allowing for easy expansion to new file formats.
- Standardized output structures specifically designed for seamless LLM processing.
SKILL.md
Loading
Related Openclaw Skills
Technical Research Skill
An advanced systematic research agent that produces multi-dimensional technical reports and architectural blueprints through real-time web intelligence.
Paper Parse
A specialized tool for converting academic PDF papers into clean markdown while extracting figures and metadata.
ClawPolicy
ClawPolicy is an explainable autonomous execution policy engine designed for low-touch, auditable agent execution and policy supervision.
Cloud Doc Intelligent Assistant
A specialized data collection tool for scraping, localizing, and monitoring changes across major cloud service provider documentation.
Perplexity Search
A web search skill that connects your AI agent to the Perplexity Sonar API for real-time, cited answers and current information.
Research Paper Reproduction Python Skill
A comprehensive AI workflow for extracting research paper content and implementing scientific methods in modular, verifiable Python code.
Featured*
