iOS has built-in functionality to allow apps to register their ability to accept and open documents of specific types. With Kagi Assistant offering best-in-class summarization and Q&A, it would be great to allow other apps to send either a URL or a document to the Kagi iOS app and have it directly input its contents into The Assistant, just waiting for the individual to select a model and enter a prompt.
For simplicity going forward, I am going to refer to all content sent over as 'the document'; however, it may be a document, image, URL, or any other format The Assistant can accept.
I see 2 different implementations here.
1 - The larger 'icon' implementation in the share menu that just sends the document with no additional context. Kagi loads this into the input field of The Assistant, and it's then up to the individual to enter a prompt, select a model, and submit.
2 - Below this, in the section that allows for sending with 'instructions,' have a 'Summarize with Kagi' button. This would send the document to Kagi and immediately trigger a summarization using whatever model Kagi thinks is best. I could also see the utility of a 'Discuss with Kagi' button, where it would also immediately send and trigger the LLM, and the response prompt would be 'What would you like to know about this document?'
