feature - image handling

[radiodee1]

Hi. I think it would be great if the program supported image processing. There are several ollama models that do visual processing, and it would be good to be able to use them. Some of them support tools. I think it would require a pass through of the image name to the underlying models. This might be easily done.

[jonigl]

Hi @radiodee1,

Thanks for the suggestion! It sounds really interesting. Could you share a bit more about how you imagine the interaction working? Also, which MCP servers are you thinking of using with vision-capable models, and how you see the image passing to the models?

This would help us better understand the use case and possibilities.

[radiodee1]

Hi. I thought at the time I wrote the suggestion - maybe incorrectly - that ollmcp used ollama under the hood. The way ollama works you just pass the path of the image to the model in the prompt. Something like the following:

>>> What is this? ./pic/figure1.png

I'm using qwen3-vl:8b for the time being.