Visual scene indexing
Jumper analyzes actual frames rather than filenames alone. Visual searches return time-bounded scenes where a described action, object, setting, composition, or emotion appears.
AI video tagging
Automated AI video tagging without manual logging
Jumper analyzes what is visible, what is said, who appears, and how each source file was recorded—automatically turning raw footage into a searchable local video index for editors and AI agents.
Definition
What AI video tagging means in Jumper
AI video tagging in Jumper means automatically creating a searchable analysis layer from what the models detect and understand; it does not mean writing keyword tags into the source files. Jumper indexes visual content, time-aligned speech, recurring people, and technical file metadata in separate local analysis files. Tag Collections are different: they are editor-created collections for deliberately saving and organizing useful segments or whole clips.
Capabilities
A richer video index for editorial work
Timestamped speech and speakers
Speech analysis creates searchable, time-aligned transcripts. Speaker detection and labels help editors locate dialogue across interviews, multicam recordings, and unscripted footage.
People and file metadata
Face analysis groups recurring people for naming and search, while metadata filters narrow footage by properties such as date, duration, codec, resolution, frame rate, camera, or lens.
Automatic intake with Watch Folders
Watch Folders can monitor selected locations and automatically analyze new media as it arrives, keeping a working footage library ready for search.
How it works
From raw footage to an editorially useful index
- 1
Choose what to analyze
Run visual, speech, and face analysis on selected media, or configure a Watch Folder for recurring intake.
- 2
Build the local index
Jumper stores compact analysis files in the chosen analysis folder. The source footage remains unchanged and does not need to be uploaded.
- 3
Search, filter, and organize
Find scenes, dialogue, and people; narrow results with metadata; then save useful segments or whole clips into personal or shared tag collections.
Technical documentation
Frequently asked questions
What does AI video tagging mean in Jumper?
It means Jumper automatically analyzes media and creates a local, searchable index for visual content, speech, people, and supported file metadata. Editors can search that AI-derived index in Jumper or let a compatible AI agent query it through Jumper.
Does Jumper provide automatic scene detection?
Jumper automatically detects the continuous time range where a visual search matches and returns it as a scene with a start and end time. This is relevance-based scene detection for search results, rather than a promise to divide every file into traditional editorial shots.
Does AI video tagging write tags into the source video files?
No. Jumper keeps its AI-generated analysis in separate local analysis files and leaves the source media unchanged.
How is AI video tagging different from Jumper Tag Collections?
AI video tagging is the automatic analysis that makes footage searchable without manual labels. Tag Collections are created intentionally by an editor to save, comment on, color-code, share, and revisit selected segments or whole clips.
Can Jumper automatically analyze newly added footage?
Yes. Watch Folders can monitor configured folders and automatically start the selected analysis when new media files appear.
Does automated video analysis require cloud uploads?
No. Jumper runs media analysis locally on macOS and Windows and stores its analysis files in the location selected by the editor.