screen-ocr.sh

#!/bin/bash

screencapture -i ~/clipboard.png

IMAGE_BASE64=$(base64 -i ~/clipboard.png | tr -d '\n')
rm ~/clipboard.png

API_KEY="GEMINI_API_KEY_HERE"
MODEL="gemini-2.0-flash"

curl -s -X POST \
  -H "Content-Type: application/json" \
  -H "X-goog-api-key: ${API_KEY}" \
  "https://generativelanguage.googleapis.com/v1beta/models/${MODEL}:generateContent" \
  -d "{
    \"contents\": [
      {
        \"parts\": [
          {\"text\": \"Extract all the text in this image. Just output the text directly without any quotes, explanation or anything else.\"},
          {
            \"inline_data\": {
              \"mime_type\": \"image/png\",
              \"data\": \"${IMAGE_BASE64}\"
            }
          }
        ]
      }
    ]
  }" | jq -r .candidates[0].content.parts[0].text | pbcopy