Semantic Decomposition · Applied to Video · Coded Prototype
Ask a question about a video. Jump to the exact moment that answers it.
Live prototype on a 74-minute NASA talk.
The Gap
The thing that makes video easy to watch is exactly what makes it hard to navigate.
Long-form video has no spatial address. A document has paragraphs, headings, a table of contents.
You can navigate it by structure. Video only has time.
Prompted Find makes the timeline navigable by meaning. You ask what you want to find.
The interface segments the video semantically, highlights relevant moments, and jumps you there.
The query returns a location, not a transcript, not a summary. A timestamp you can trust.
The Method
Semantic Decomposition: break long-form content into labeled, navigable semantic units,
so a question can return a location instead of a summary.
Prompted Find is this method applied to video. The same decomposition works on any surface that has
length but no structure: prose, podcasts, research trails.
Why this pattern matters beyond video. The same logic applies to any long-form content without navigable structure: podcasts, recorded meetings, legal documents, research papers. Prompted Find is a prototype of a pattern: ask a question, get a location. The medium changes. The interaction stays the same.
Decisions
01
Timestamps instead of text
“Search as navigation.”
The query moves you inside the video, to the moment where the answer lives.
02
Segments instead of search results
“The timeline is the interface.”
Labeled segments on the timeline show the shape of the answer before you click.
03
Follow-ups instead of starting over
“Context persists.”
The question panel stays open beside the video, so you never lose your place.
The Interaction
What I'd Change