"author:ergou-yu" matching MCP servers:

Enables text-only LLMs to perceive images entirely on-device, providing vision capabilities like image description, OCR, table extraction, UI analysis, and region focusing without any cloud APIs or API keys.