Overview

Jumper is a machine-learning search engine for video footage, built for editors. It lets you find shots, scenes, spoken words and people across your footage library using plain keywords, without tagging clips first. A key point is that it runs on your own device: footage is analysed locally and does not need to be uploaded to the cloud, and the tool can work offline.

Where it works

Jumper is available as a standalone app and as plugins for Final Cut Pro, Adobe Premiere Pro, DaVinci Resolve and Avid, so editors can search without leaving their editing software. The company says it is used by everyone from solo creators to large studios, and describes it as agent-ready, meaning an AI agent can use it to look for shots.

Search features

Visual search returns relevant clips for a keyword such as a type of shot or object. Spoken content search finds where words or phrases are said. Match Similar shows visually similar moments to one you have found, which helps when building sequences or finding alternative takes. Match Source Monitor searches for clips that match the frame currently on your source monitor. Face recognition finds every appearance of a person across a project, which is useful for documentaries, series and archive work. Testimonials on the site come from editors, video specialists and producers who say it saves them time.

Pricing

Jumper offers a free trial with no credit card required. Paid plans follow the trial; check the website for current prices and licence options.

Who it suits

Jumper suits video editors, documentary makers, post-production teams and content creators who work with large amounts of footage and spend a lot of time logging and searching for clips.

Things to consider

Local processing keeps footage private but uses your computer's resources, so indexing a large library may take time and storage. Check which versions of your editing software are supported and how the index is shared with collaborators. Face recognition involves personal data, so consider consent and local privacy rules for footage of identifiable people. Test search accuracy on your own material, such as interviews in other languages or low-light shots.