Workflows
Video Library Organization: From Folder Chaos to Instant Search
Organization and search are different problems. Content search finds footage in a library that was never reorganized, and lets you fix the folders on your own schedule.
Most editors have a drive from a few years back with three overlapping naming conventions, a NAS volume where four people dumped footage with no agreed structure, or a drawer of external drives labelled in marker. The gap between how the footage is organised and how it should be grows with every project, until reorganising is too big a job to start.
Content search does not need the folders fixed first.
Common organisation problems
Duplicate filenames. Camera filenames like A001_C003_0101.braw and MVI_4721.MP4 repeat across projects. A filename search returns dozens of unrelated clips. Renaming thousands of files after the fact breaks project references.
Scattered drives. Footage sits on internal drives, external SSDs, USB hard drives and NAS volumes. Finding a clip means remembering which drive it is on, and whether that drive is plugged in.
Lost project folders. "We shot this two years ago for the Henderson project." The Henderson folder was renamed, or sits on a drive that was reformatted, or is nested three levels inside someone else's folder.
Inconsistent naming. Some folders are organised by date, some by client, some by project. Browsing requires knowing which scheme was in use in which year.
Renders mixed with originals. Exports, proxies, render caches and camera originals share a folder. A search returns every derivative alongside the source.
Organisation and search are different problems
Folder organisation answers "where did I put this?" Content search answers "which footage contains this?"
A well-organised library with descriptive folder names still cannot answer "which clip has Sarah talking about the product launch." You can browse to the right project folder and the right shoot day. You still scrub through the clips.
A disorganised library with content search answers that question directly. The search reads the words spoken, the objects visible, the people on screen and the caption for each scene. It does not read the filename, the folder or the drive.
Both are worth having. Organisation helps you browse. Search helps you find.
Searching an unorganised library
FrameQuery extracts four kinds of data from each clip:
- Transcripts, timestamped line by line.
- Object detection results for items visible in each frame.
- Scene descriptions in natural language.
- Face and voice clusters for the people in the library.
The data goes into a local Tantivy search index. A clip named A001_C003_0101.braw in an unlabelled folder is as searchable as 2025-03-15_Henderson_Interview_CamA_Take3.mov in a tidy directory.
Point FrameQuery at the drives as they are, process the footage, and search. Nothing is renamed or moved.
What happens to the backlog
The Henderson clips split across two drives and a NAS are in one index. Footage with camera filenames is searchable by dialogue, visible content and the people in it. The drive with three naming conventions is indexed the same way as everything else.
The folders stay as messy as they were. You can reorganise later, or not.
Steps
Start with search. Add existing drives and folders to FrameQuery as they are and process them.
Exclude what you do not need. Use folder exclusion rules to skip render caches, proxy folders and export directories.
Organise going forward. Pick a convention for new projects, such as Year/ProjectName/CameraOriginals. Do not apply it to old projects. They are already searchable.
Use projects for grouping. Assign clips to projects inside FrameQuery. That layer sits on top of the file system without changing it. Search within a project for focused results or across the library for a wider net.
Leave the archive alone. Five-year-old footage with bad filenames is searchable now. Spend the time on the current edit.
Download FrameQuery to make your footage library searchable.