Image Analysis & OCR via Crazyrouter#

Analyze images, extract text, and describe visual content using vision models through Crazyrouter.

Script Directory#

Agent Execution:

1. SKILL_DIR = this SKILL.md file's directory
2. Script path = ${SKILL_DIR}/scripts/main.mjs

Step 0: Check API Key ⛔ BLOCKING#

bash

[ -n "${CRAZYROUTER_API_KEY}" ] && echo "key_present" || echo "not_set"

ResultActionkey_presentContinuenot_setAsk user to set CRAZYROUTER_API_KEY. Get key at https://crazyrouter.com

Usage#

bash

# Describe image
node ${SKILL_DIR}/scripts/main.mjs --image photo.jpg --prompt "Describe this image"

# Extract text (OCR)
node ${SKILL_DIR}/scripts/main.mjs --image receipt.png --prompt "Extract all text from this image"

# Analyze screenshot
node ${SKILL_DIR}/scripts/main.mjs --image screenshot.png --prompt "What errors are shown in this screenshot?"

# Use specific model
node ${SKILL_DIR}/scripts/main.mjs --image chart.png --prompt "Analyze this chart" --model gemini-2.5-flash

# Save output
node ${SKILL_DIR}/scripts/main.mjs --image document.jpg --prompt "Extract text" --output text.md

Options#

OptionDescriptionDefault--image <path>Image file (required)—--prompt <text>What to do with the imageDescribe this image in detail``--model <id>Vision modelgpt-4o``--output <file>Save to filestdout