Dip — give LLMs eyes
Cursor AI Hackathon
1st · Google DeepMind Gemini Track$10,000 Gemini credits
Streams your screen into the model in real time, so you stop describing what is on it.
Everyone at the table had felt the same thing: explaining to the LLM what we see on screen, because it cannot see. Dip captures the screen and feeds it into Gemini as live context for pair-programming and debugging. A lightweight client captures, a Python backend streams. Best use of the Google DeepMind Gemini API out of 400+ participants.
PythonGemini APIScreen capture
