Search your footage the way you remember it

The problem it solves

Finding one moment in four hours is a scrubbing job, and scrubbing is how an afternoon disappears. Filenames tell you nothing and folders tell you less.

What it needs first

• Ingested footage. The search runs over each clip's pre-built visual index and transcript, so both are built at ingest rather than at search time.

How it works

1. 1 Ask in plain wordsDescribe the moment the way you would to a person, not the way a filename would.
2. 2 It matches on both indexesWhat was said and what is visible, so a silent shot is as findable as a spoken line.
3. 3 Collect into a binHits gather into a named bin you can hand over, which is a different deliverable from a four-hour card and a hopeful message.

What you get back

Matching clips for a plain-language query

Ask for 'the rooftop argument' or 'coffee shops' and get clips back, matched on what was said AND what is on screen.

Why each clip matched

Said or on-screen, per hit, so you can tell a spoken mention from a visual appearance.

A collected bin

Matches can be grouped into a named workspace folder to hand to an editor, instead of a four-hour card and a description.

What it will not do

• It searches meaning, not filenames. That is the point: it works on cards you never labelled.
• Filing is careful about clips Organize already placed. Unfiled clips are moved into the bin; already-filed ones are copied, so a search never undoes your beat structure.
• Recall depends on the visual index. Something on screen for half a second and never mentioned aloud is the weak case.

What it buys you

• A moment becomes something you search for instead of hunt for.
• Silent B-roll is searchable, because every clip is described and not only transcribed.
• The result is a bin, so the search survives past the moment you ran it.

Questions