A powerful tool for converting complex PDFs and scanned documents into structured Markdown and JSON using the Mistral OCR API.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install extracting-mistral-ocr
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 extracting-mistral-ocr 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 Mistral OCR PDF Extraction?
This skill leverages the Mistral OCR API to provide high-fidelity document understanding, capable of processing both digital and scanned PDFs. It is designed to bridge the gap between unstructured documents and LLM-ready data, making it an essential component within the Openclaw Skills ecosystem for developers building RAG systems or document processing pipelines.
By automating the extraction of text, tables, and images, it ensures that your AI agents have access to the full context of any document. The skill handles the complexity of file uploads, API communication, and output organization, providing a developer-friendly interface for document intelligence.
Mistral OCR PDF Extraction Use Cases
- Converting scanned PDF documents into searchable Markdown for RAG applications.
- Extracting structured data like invoice totals or contract dates using custom annotation prompts.
- Parsing complex multi-page reports while preserving table structures and visual elements.
- Automating document ingestion pipelines that require high-accuracy optical character recognition.
How Mistral OCR PDF Extraction Works
- The script identifies the input source, supporting both local file paths and public URLs.
- Local files are automatically uploaded to the Mistral Files API to generate a unique file ID for processing.
- The OCR request is sent to Mistral with specific configuration parameters for table formatting and image extraction.
- The API processes the document and returns a detailed response including page-level markdown, metadata, and optional annotations.
- The skill organizes the output into a structured directory containing the full text, individual pages, and any extracted assets.
Mistral OCR PDF Extraction Setup
Ensure you have an active Mistral API key and Python 3.9+ installed with the mistralai package. This tool is a core part of the Openclaw Skills collection.
# Set your environment variable
export MISTRAL_API_KEY='your_api_key_here'
# Run the extraction script
python scripts/mistral_ocr_extract.py --input path/to/file.pdf --out out/ocr
Mistral OCR PDF Extraction Data Schema & Taxonomy
The skill generates a deterministic output folder with the following structure to ensure compatibility with downstream Openclaw Skills:
| File/Folder | Description |
|---|---|
| combined.md | The entire document concatenated into a single Markdown file. |
| pages/ | Individual Markdown files for each page (e.g., page-000.md). |
| raw_response.json | The complete, unedited JSON response from the Mistral OCR API. |
| images/ | Decoded images and diagrams extracted from the PDF if requested. |
| tables/ | Extracted tables saved as separate files if HTML format is requested. |
Mistral OCR PDF Extraction Advanced Features
- Structured Extraction: Pass custom annotation prompts to extract specific JSON fields from the document level.
- High-Fidelity Tables: Support for HTML table formatting to improve parsing accuracy for complex data sets.
- Asset Decoding: Automatic handling and decoding of base64-encoded images for visual document context.
- Selective Processing: Option to process specific page ranges to optimize token usage and processing speed.
SKILL.md
Loading
Related Openclaw Skills
App Store Readiness Audit
An automated compliance auditor for iOS repositories ensuring App Store and TestFlight readiness.
AI Codebase Deep Modules
A specialized skill for refactoring software architectures into deep, AI-navigable modules with enforced boundaries and fast feedback loops.
Readwise MCP (mcporter)
Connect your Readwise and Reader library to AI agents using the Model Context Protocol for semantic search and document management.
Nexus Nova
A sophisticated personality architecture that transforms AI agents from simple scripts into autonomous entities with genuine identity and emotional depth.
Fabric API Skill
A cross-platform skill for Openclaw Skills that enables AI agents to create, search, and organize content within a Fabric workspace using the official HTTP API.
iOS Simulator Automation
An advanced CLI wrapper that enables AI agents to programmatically control iOS simulators, manage app lifecycles, and perform accessibility-based UI navigation.
Featured*
