Feature Requests

Would love to be able to have a team access and search through clips as well. Not sure how this would work, but even if I could just share a "jumper project file" with someone so they can search and mark timecodes, that would be so helpful. They don't need to have the footage itself, just the transcriptions and searching.

As a user, I want to filter search results by video orientation (vertical/horizontal) When searching for footage, I want to apply filters that allow me to see only vertical or horizontal videos. This will help me quickly find content that fits my project’s aspect ratio without sifting through irrelevant results.

If you have some footage with visual elements or concepts that is difficult to find accurately, for example if the search results are consistently confusing or performing poorly for one "class" or type of shot/frame, we could fine-tune a model if we are supplied the training data - i.e. the frames and what they are (label, name, whatever the variations may be). Kind of a labor intensive job for the user to provide all the required frames (or exact timestamps where to find frames at least, so we can extract them with a script) To clarify - we need a dataset of examples, not just a single image or two.

"As a user, I want to see only one search result per video file" When searching for footage, I often see multiple scenes from the same video file, which can clutter the results. I want an option to display only one result per video, so I can decide whether to explore more fragments from that file if needed.

To let you search transcriptions not by finding exact matches to your searches, but rather find them by semantic meaning. So for example if someone said "vehicle" and you searched for "car", you would still find the transcription result where someone said "car", even though none of the letters in the word "car" matches "vehicle".

Enable prompt based textal conversation with your media. Using the analysis and the transcriptions, interact with your media via text in arbitrary ways (e.g., ask questions about it to gain insights)

Enable creating rough cuts / selects on a timeline by natural language prompting. So "a timeline of 10 shots where people get increasingly more frustrated" (need a video editor to give me a better example here 😄). See attached screenshot of a mockup example GUI that @tontonius created.

The option to export 10-20 top keywords to the relevant FCPX clip's keywords. This would facilitate the use of smart collections. Also, being able to add metadata directly to the file would facilitate archiving footage.