If I may, are there plans for visual input support?

We have internal prototypes, so please reach out to hello@inceptionlabs.ai if you're interested in early access