Show HN: I taught GPT-OSS-120B to see using Google Lens and OpenCV
Summary
This Hacker News Show HN post describes teaching GPT-OSS-120B, a text-only large language model, to perform vision tasks using Google Lens and OpenCV. It details an MCP server setup that provides real Google search and vision capabilities without API keys, with Google Lens-based object detection and cropping to identify objects. The discussion includes notes on potential TOS concerns and reliability, and provides GitHub and PyPI links for the project.