WiseOCR Document Recognition to Markdown

The Problem

• Text in scanned pages, screenshots, or a single PDF needs to become editable Markdown, but local OCR can be heavy and cloud APIs can require manual request code.
• This skill targets a single PDF or image file, calls WiseDiag OCR to produce Markdown, and writes the result to a fixed workspace directory.

How It Works

• Supported formats: PDF, jpg, jpeg, png, webp, gif, bmp, and tiff.
• A WiseDiag API key must be configured; the skill reads it from the environment and avoids direct hand-written HTTP calls.
• Key parameters include the input file, output name stem, output directory, and PDF rendering DPI; the default DPI is 200 and the supported range is 72-600.
• If the input file is copied, renamed, or moved to a temporary path, pass the original filename stem explicitly so the output name remains readable.
• The result is saved as a Markdown file, defaulting to ~/.openclaw/workspace/WiseOCR/{name}.md, with no extra manual saving step.

Boundaries

• Files are uploaded to WiseDiag cloud servers for processing, so this is not suitable for sensitive, confidential, or compliance-heavy documents.
• The documented scope is a single file, not batch directories, complex layout reconstruction, or local private OCR.
• Cloud processing makes the skill dependent on network access, API availability, and third-party data handling.