paddleocr-text-recognition

Turn images and scanned PDFs into complete text.

optical character recognition

What it does

Extracts machine-readable, line-level text from images and scanned PDFs.

When to use it

Use it to extract line- or box-level text from image and PDF sources.

Give it an image or PDF URL or local path; it returns the complete recognized text.

What you provide

This skill

PaddleOCR OCR API

Write

Sends data to PaddleOCR

httpx

Execute

Resolves the httpx dependency

system temporary directory

Write

Saves raw OCR JSON

Requires Python 3.9 or newer.

Requires uv to run the bundled script and resolve its dependencies.

The default script requires a configured PaddleOCR OCR API endpoint ending with /ocr.

The default script reads the PaddleOCR access token from the environment.