Does anyone know of a plugin that actually works, which will allow a block of text to be highlighted and used as a “prompt” to generate an image using AI, say Claude or CGPT, that can be inserted within a note?
I run a busy newsletter, and the one thing I cannot do within Obsidian is write the newsletter then generate images “in-house.”
I must have tried every one out there, and they all fail, usually due to a lack of a DALL-E option.
Anyone? (I don’t mind paying if it works.)
Not an image maker, but if you install Agent Client or its fork Agent Console, you can add custom agents, e.g. Hermes Agent in them, and with Hermes you can do it, according to this Reddit thread, for instance.
The image may not be rendered in the chat windows, but the important thing is to have the result on the disk.
I just tried it. Deepseek V4 Flash (free) model running through custom Hermes Agent via its default acp bridge in Agent Client accessing, managing and prompting my dumbed-down Comfy-UI with the server running (I showed it the terminal output of the server running).
Hermes default install has built in ComfyUI skill.
Hermes Agent is good because you can custom create a skill that follows the tool handling that works and can make memories on best practices paying dividends on its own.
I did not try a cloud model but would probably easy to set up. This dumbed down Comfy is only good for mediocre looking pics, BTW.
Not the solution i was looking for but thanks.