Giving your local AI system eyes

Hi everything,

Galvin here, cofounder of Lauretta and one of the developers of Hearthlight.

We’re building what we call AI for the physical world. A local system that connects cameras and videos to LLMs and allows it to run a library of triggers. You can connect it to Openclaw and give it eyes, or to robot systems to automate when they do things.

One thing about LLM linked systems is that you can change what it is looking for (like a pose, or behavior or an object) via modifying the prompt, no need to retrain anything.

I'd love to get some feedback on the system, what you think you might do with it and what we can improve.