ocr

name	ocr
description	Extract text from images and scanned PDFs using local OCR (Tesseract). Keyless, offline. Use when asked to "read text from this image", "OCR this", "extract text from a screenshot/scan", or given an image with text.
allowed-tools	["@coder","Bash","@read"]
triggers	["ocr","read text from image","extract text from image","text from screenshot","scan to text","ler texto da imagem","extrair texto da imagem","texto do print"]

OCR (image / scan → text)

Local, offline OCR via Tesseract — no API, no upload.

Step 0 — Detect

• macOS / Linux: command -v tesseract pdftoppm 2>/dev/null; tesseract --list-langs 2>/dev/null
• Windows: Get-Command tesseract -ErrorAction SilentlyContinue

Install: brew install tesseract / apt install tesseract-ocr / choco install tesseract. Language packs: tesseract-ocr-por (Portuguese), -eng (English), etc.

Image → text

tesseract input.png stdout -l eng
tesseract input.jpg stdout -l por+eng     # multi-language

Write to a file with tesseract input.png out → out.txt, then read with @read.

Scanned PDF → text

Rasterize pages first, then OCR each:

pdftoppm -png -r 300 scan.pdf page
for f in page-*.png; do tesseract "$f" stdout -l por+eng; done > scan.txt

(On Windows, loop with PowerShell Get-ChildItem page-*.png | % { tesseract $_.Name stdout -l por+eng }.)

Tips for accuracy

• Use 300 DPI for scans; higher helps small text.
• -l must match the document language — pick por for Portuguese, or it mangles accents.
• For skewed/noisy scans, mention that pre-processing (deskew/threshold via ImageMagick) improves results.

Rules

• Choose the language pack from the document, not the UI locale.
• Present the extracted text, then offer to summarize/translate/feed it to another skill.