Whats the status on multimodal documents?

Are there any plans to make it possible to upload videos/img and use that to query?

Certainly part of the vision but not part of the roadmap in the next 6 months unfortunately.