I generally prefer to use Gemini models but it seems they are not yet available when setting up a task template.
Could you please enable selecting not just “Task”, “Deep” etc. but also Gemini? I can select Gemini at the bottom but only if no template is set, and in this scenario it seems tools are disabled by default. Users should be able to enable/disable tools when not using a prompt template, currently it seems to only be possible when using a template but in this case the model is locked.
In general, it would be nice to have more control over the model selection. Ideally, it would even be possible to set custom models via an API key or log in via oAuth.
Either way, please offer gemini-3-flash-preview as soon as possible, since this is a huge improvement over gemini 2.5 flash at marginally higher cost.
I appreciate you looking into these suggestions. Your work integrating AI is extremely appreciated - it is already adding a ton of value!
Yes, that is understood. However, then the model only has access to Read tools. It is no longer able to edit and make changes to the mind map. See conversation screen shot above.
To reproduce this, you can switch the model to Gemini and tell it to edit a card. This does not work.
I finally got it. Setting a model in the template locks it for the rest of the conversation.
Thank you very much for your patience
Please remember to consider my other suggestions:
offer gemini 3 flash preview
usage via API key or oAuth token
PS: I did get Gemini to implement edits in the map, it is possible. But: For some reason sometimes it still insists that it can only use Read tools. Setting the model to Task works more reliably.
Since the task model is specifically optimized for task processing, other models sometimes output content with incorrect formatting, causing the software to fail to recognize these modifications
