Jumper relies on advanced machine learning models to find things in your footage.

​
Search by text

The most common way to search is by text. Simply type whatever you’re looking for in the Search box and click Search. SCR-20250511-jegb.avif Searching by text works best in english but if you have analyzed your footage with a multilingual model, Jumper will find what you’re looking for in ~80 languages. For example you can search for white cat, gato blanco or 白い猫 and most likely get the same results
SCR-20250511-jesb.avif

With a multilingual model, you can search in other languages.

​
Search by image

Another way is to use an image to search. When searching by image, Jumper will “look” at the image provided and find scenes in your footage with visual similarity to the search image. SCR-20250511-jgwp.jpeg This can be especially useful for finding more footage of the same person, location or setting. There are four ways to search by image:
1

Match Source/Event Monitor

Takes the frame currently visible in the Source Monitor and performs a search with it.
2

Match Timeline

Takes the frame currently visible in the Timeline Monitor and performs a search with it.
3

Match Pasteboard

Takes the image currently in your pasteboard and performs a search with it.
4

Find Similar

Takes the image of a search result thumbnail and performs a search with it.

​
Results

After performing a search Jumper returns results. A result is a scene , i.e. a segment of a file with a start and an end. scenes-in-files.avif