Paste image as alternative to browsing for file

[glinkot]

When asking chatGPT vision (or other model supporting images) to look at something, it would be great to take a screenshot and just hit ctrl-v while entering the prompt, rather than saving the file, choosing a name, then browsing again to find it.

[kurianbenoy]

Ollama is already supported right?

[HarshShah0203]

how to use Ollama for this

[vishnusureshperumbavoor]

I selected mistral and I got this error

https://github.com/stitionai/devika/issues/25

[rajakumar05032000]

#75 Added LlamaCpp which will allow to load local models, It eliminates the need for separate local LLM inference server